☆ FrugalNeRF: Fast Convergence for Few-shot Novel View Synthesis without Learned Priors

Chin-Yang Lin, Chung-Ho Wu, Chang-Han Yeh, Shih-Han Yen, Cheng Sun, Yu-Lun Liu

Neural Radiance Fields (NeRF) face significant challenges in few-shot scenarios, primarily due to overfitting and long training times for high-fidelity rendering. Existing methods, such as FreeNeRF and SparseNeRF, use frequency regularization or pre-trained priors but struggle with complex scheduling and bias. We introduce FrugalNeRF, a novel few-shot NeRF framework that leverages weight-sharing voxels across multiple scales to efficiently represent scene details. Our key contribution is a cross-scale geometric adaptation scheme that selects pseudo ground truth depth based on reprojection errors across scales. This guides training without relying on externally learned priors, enabling full utilization of the training data. It can also integrate pre-trained priors, enhancing quality without slowing convergence. Experiments on LLFF, DTU, and RealEstate-10K show that FrugalNeRF outperforms other few-shot NeRF methods while significantly reducing training time, making it a practical solution for efficient and accurate 3D scene reconstruction.

comment: Project page: https://linjohnss.github.io/frugalnerf/

☆ MvDrag3D: Drag-based Creative 3D Editing via Multi-view Generation-Reconstruction Priors

Honghua Chen, Yushi Lan, Yongwei Chen, Yifan Zhou, Xingang Pan

Drag-based editing has become popular in 2D content creation, driven by the capabilities of image generative models. However, extending this technique to 3D remains a challenge. Existing 3D drag-based editing methods, whether employing explicit spatial transformations or relying on implicit latent optimization within limited-capacity 3D generative models, fall short in handling significant topology changes or generating new textures across diverse object categories. To overcome these limitations, we introduce MVDrag3D, a novel framework for more flexible and creative drag-based 3D editing that leverages multi-view generation and reconstruction priors. At the core of our approach is the usage of a multi-view diffusion model as a strong generative prior to perform consistent drag editing over multiple rendered views, which is followed by a reconstruction model that reconstructs 3D Gaussians of the edited object. While the initial 3D Gaussians may suffer from misalignment between different views, we address this via view-specific deformation networks that adjust the position of Gaussians to be well aligned. In addition, we propose a multi-view score function that distills generative priors from multiple views to further enhance the view consistency and visual quality. Extensive experiments demonstrate that MVDrag3D provides a precise, generative, and flexible solution for 3D drag-based editing, supporting more versatile editing effects across various object categories and 3D representations.

comment: 16 pages, 10 figures, conference

☆ SAM2Long: Enhancing SAM 2 for Long Video Segmentation with a Training-Free Memory Tree

Shuangrui Ding, Rui Qian, Xiaoyi Dong, Pan Zhang, Yuhang Zang, Yuhang Cao, Yuwei Guo, Dahua Lin, Jiaqi Wang

The Segment Anything Model 2 (SAM 2) has emerged as a powerful foundation model for object segmentation in both images and videos, paving the way for various downstream video applications. The crucial design of SAM 2 for video segmentation is its memory module, which prompts object-aware memories from previous frames for current frame prediction. However, its greedy-selection memory design suffers from the "error accumulation" problem, where an errored or missed mask will cascade and influence the segmentation of the subsequent frames, which limits the performance of SAM 2 toward complex long-term videos. To this end, we introduce SAM2Long, an improved training-free video object segmentation strategy, which considers the segmentation uncertainty within each frame and chooses the video-level optimal results from multiple segmentation pathways in a constrained tree search manner. In practice, we maintain a fixed number of segmentation pathways throughout the video. For each frame, multiple masks are proposed based on the existing pathways, creating various candidate branches. We then select the same fixed number of branches with higher cumulative scores as the new pathways for the next frame. After processing the final frame, the pathway with the highest cumulative score is chosen as the final segmentation result. Benefiting from its heuristic search design, SAM2Long is robust toward occlusions and object reappearances, and can effectively segment and track objects for complex long-term videos. Notably, SAM2Long achieves an average improvement of 3.0 points across all 24 head-to-head comparisons, with gains of up to 5.3 points in J&F on long-term video object segmentation benchmarks such as SA-V and LVOS. The code is released at https://github.com/Mark12Ding/SAM2Long.

comment: Project page: https://mark12ding.github.io/project/SAM2Long/

☆ xGen-MM-Vid (BLIP-3-Video): You Only Need 32 Tokens to Represent a Video Even in VLMs

Michael S. Ryoo, Honglu Zhou, Shrikant Kendre, Can Qin, Le Xue, Manli Shu, Silvio Savarese, Ran Xu, Caiming Xiong, Juan Carlos Niebles

We present xGen-MM-Vid (BLIP-3-Video): a multimodal language model for videos, particularly designed to efficiently capture temporal information over multiple frames. BLIP-3-Video takes advantage of the 'temporal encoder' in addition to the conventional visual tokenizer, which maps a sequence of tokens over multiple frames into a compact set of visual tokens. This enables BLIP3-Video to use much fewer visual tokens than its competing models (e.g., 32 vs. 4608 tokens). We explore different types of temporal encoders, including learnable spatio-temporal pooling as well as sequential models like Token Turing Machines. We experimentally confirm that BLIP-3-Video obtains video question-answering accuracies comparable to much larger state-of-the-art models (e.g., 34B), while being much smaller (i.e., 4B) and more efficient by using fewer visual tokens. The project website is at https://www.salesforceairesearch.com/opensource/xGen-MM-Vid/index.html

☆ 3DGS-Enhancer: Enhancing Unbounded 3D Gaussian Splatting with View-consistent 2D Diffusion Priors NeurIPS 2024

Xi Liu, Chaoyi Zhou, Siyu Huang

Novel-view synthesis aims to generate novel views of a scene from multiple input images or videos, and recent advancements like 3D Gaussian splatting (3DGS) have achieved notable success in producing photorealistic renderings with efficient pipelines. However, generating high-quality novel views under challenging settings, such as sparse input views, remains difficult due to insufficient information in under-sampled areas, often resulting in noticeable artifacts. This paper presents 3DGS-Enhancer, a novel pipeline for enhancing the representation quality of 3DGS representations. We leverage 2D video diffusion priors to address the challenging 3D view consistency problem, reformulating it as achieving temporal consistency within a video generation process. 3DGS-Enhancer restores view-consistent latent features of rendered novel views and integrates them with the input views through a spatial-temporal decoder. The enhanced views are then used to fine-tune the initial 3DGS model, significantly improving its rendering performance. Extensive experiments on large-scale datasets of unbounded scenes demonstrate that 3DGS-Enhancer yields superior reconstruction performance and high-fidelity rendering results compared to state-of-the-art methods. The project webpage is https://xiliu8006.github.io/3DGS-Enhancer-project .

comment: Accepted by NeurIPS 2024 Spotlight

☆ Mini-InternVL: A Flexible-Transfer Pocket Multimodal Model with 5% Parameters and 90% Performance

Zhangwei Gao, Zhe Chen, Erfei Cui, Yiming Ren, Weiyun Wang, Jinguo Zhu, Hao Tian, Shenglong Ye, Junjun He, Xizhou Zhu, Lewei Lu, Tong Lu, Yu Qiao, Jifeng Dai, Wenhai Wang

Multimodal large language models (MLLMs) have demonstrated impressive performance in vision-language tasks across a broad spectrum of domains. However, the large model scale and associated high computational costs pose significant challenges for training and deploying MLLMs on consumer-grade GPUs or edge devices, thereby hindering their widespread application. In this work, we introduce Mini-InternVL, a series of MLLMs with parameters ranging from 1B to 4B, which achieves 90% of the performance with only 5% of the parameters. This significant improvement in efficiency and effectiveness makes our models more accessible and applicable in various real-world scenarios. To further promote the adoption of our models, we develop a unified adaptation framework for Mini-InternVL, which enables our models to transfer and outperform specialized models in downstream tasks, including autonomous driving, medical images, and remote sensing. We believe that our study can provide valuable insights and resources to advance the development of efficient and effective MLLMs. Code is available at https://github.com/OpenGVLab/InternVL.

comment: Technical report

☆ Agent-to-Sim: Learning Interactive Behavior Models from Casual Longitudinal Videos

Gengshan Yang, Andrea Bajcsy, Shunsuke Saito, Angjoo Kanazawa

We present Agent-to-Sim (ATS), a framework for learning interactive behavior models of 3D agents from casual longitudinal video collections. Different from prior works that rely on marker-based tracking and multiview cameras, ATS learns natural behaviors of animal and human agents non-invasively through video observations recorded over a long time-span (e.g., a month) in a single environment. Modeling 3D behavior of an agent requires persistent 3D tracking (e.g., knowing which point corresponds to which) over a long time period. To obtain such data, we develop a coarse-to-fine registration method that tracks the agent and the camera over time through a canonical 3D space, resulting in a complete and persistent spacetime 4D representation. We then train a generative model of agent behaviors using paired data of perception and motion of an agent queried from the 4D reconstruction. ATS enables real-to-sim transfer from video recordings of an agent to an interactive behavior simulator. We demonstrate results on pets (e.g., cat, dog, bunny) and human given monocular RGBD videos captured by a smartphone.

comment: Project page: https://gengshan-y.github.io/agent2sim-www/

☆ Elucidating the design space of language models for image generation

Xuantong Liu, Shaozhe Hao, Xianbiao Qi, Tianyang Hu, Jun Wang, Rong Xiao, Yuan Yao

The success of autoregressive (AR) language models in text generation has inspired the computer vision community to adopt Large Language Models (LLMs) for image generation. However, considering the essential differences between text and image modalities, the design space of language models for image generation remains underexplored. We observe that image tokens exhibit greater randomness compared to text tokens, which presents challenges when training with token prediction. Nevertheless, AR models demonstrate their potential by effectively learning patterns even from a seemingly suboptimal optimization problem. Our analysis also reveals that while all models successfully grasp the importance of local information in image generation, smaller models struggle to capture the global context. In contrast, larger models showcase improved capabilities in this area, helping to explain the performance gains achieved when scaling up model size. We further elucidate the design space of language models for vision generation, including tokenizer choice, model choice, model scalability, vocabulary design, and sampling strategy through extensive comparative experiments. Our work is the first to analyze the optimization behavior of language models in vision generation, and we believe it can inspire more effective designs when applying LMs to other domains. Finally, our elucidated language model for image generation, termed as ELM, achieves state-of-the-art performance on the ImageNet 256*256 benchmark. The code is available at https://github.com/Pepperlll/LMforImageGeneration.git.

comment: Project page: https://pepper-lll.github.io/LMforImageGeneration/

☆ Revisiting Deep Feature Reconstruction for Logical and Structural Industrial Anomaly Detection

Sukanya Patra, Souhaib Ben Taieb

Industrial anomaly detection is crucial for quality control and predictive maintenance, but it presents challenges due to limited training data, diverse anomaly types, and external factors that alter object appearances. Existing methods commonly detect structural anomalies, such as dents and scratches, by leveraging multi-scale features from image patches extracted through deep pre-trained networks. However, significant memory and computational demands often limit their practical application. Additionally, detecting logical anomalies-such as images with missing or excess elements-requires an understanding of spatial relationships that traditional patch-based methods fail to capture. In this work, we address these limitations by focusing on Deep Feature Reconstruction (DFR), a memory- and compute-efficient approach for detecting structural anomalies. We further enhance DFR into a unified framework, called ULSAD, which is capable of detecting both structural and logical anomalies. Specifically, we refine the DFR training objective to improve performance in structural anomaly detection, while introducing an attention-based loss mechanism using a global autoencoder-like network to handle logical anomaly detection. Our empirical evaluation across five benchmark datasets demonstrates the performance of ULSAD in detecting and localizing both structural and logical anomalies, outperforming eight state-of-the-art methods. An extensive ablation study further highlights the contribution of each component to the overall performance improvement. Our code is available at https://github.com/sukanyapatra1997/ULSAD-2024.git

comment: Accepted in Transactions on Machine Learning Research (TMLR). Link to OpenReview: https://openreview.net/forum?id=kdTC4ktHPD

☆ MoRE: Multi-Modal Contrastive Pre-training with Transformers on X-Rays, ECGs, and Diagnostic Report

Samrajya Thapa, Koushik Howlader, Subhankar Bhattacharjee, Wei le

In this paper, we introduce a novel Multi-Modal Contrastive Pre-training Framework that synergistically combines X-rays, electrocardiograms (ECGs), and radiology/cardiology reports. Our approach leverages transformers to encode these diverse modalities into a unified representation space, aiming to enhance diagnostic accuracy and facilitate comprehensive patient assessments. We utilize LoRA-Peft to significantly reduce trainable parameters in the LLM and incorporate recent linear attention dropping strategy in the Vision Transformer(ViT) for smoother attention. Furthermore, we provide novel multimodal attention explanations and retrieval for our model. To the best of our knowledge, we are the first to propose an integrated model that combines X-ray, ECG, and Radiology/Cardiology Report with this approach. By utilizing contrastive loss, MoRE effectively aligns modality-specific features into a coherent embedding, which supports various downstream tasks such as zero-shot classification and multimodal retrieval. Employing our proposed methodology, we achieve state-of-the-art (SOTA) on the Mimic-IV, CheXpert, Edema Severity, and PtbXl downstream datasets, surpassing existing multimodal approaches. Our proposed framework shows significant improvements in capturing intricate inter-modal relationships and its robustness in medical diagnosis that establishes a framework for future research in multimodal learning in the healthcare sector.

comment: 10 pages, 5 figures, 9 tables. Supplementary detail in Appendix. Code made available in Github for reproducibility

☆ Deep Radiomics Detection of Clinically Significant Prostate Cancer on Multicenter MRI: Initial Comparison to PI-RADS Assessment

G. A. Nketiah, M. R. Sunoqrot, E. Sandsmark, S. Langørgen, K. M. Selnæs, H. Bertilsson, M. Elschot, T. F. Bathen

Objective: To develop and evaluate a deep radiomics model for clinically significant prostate cancer (csPCa, grade group >= 2) detection and compare its performance to Prostate Imaging Reporting and Data System (PI-RADS) assessment in a multicenter cohort. Materials and Methods: This retrospective study analyzed biparametric (T2W and DW) prostate MRI sequences of 615 patients (mean age, 63.1 +/- 7 years) from four datasets acquired between 2010 and 2020: PROSTATEx challenge, Prostate158 challenge, PCaMAP trial, and an in-house (NTNU/St. Olavs Hospital) dataset. With expert annotations as ground truth, a deep radiomics model was trained, including nnU-Net segmentation of the prostate gland, voxel-wise radiomic feature extraction, extreme gradient boost classification, and post-processing of tumor probability maps into csPCa detection maps. Training involved 5-fold cross-validation using the PROSTATEx (n=199), Prostate158 (n=138), and PCaMAP (n=78) datasets, and testing on the in-house (n=200) dataset. Patient- and lesion-level performance were compared to PI-RADS using area under ROC curve (AUROC [95% CI]), sensitivity, and specificity analysis. Results: On the test data, the radiologist achieved a patient-level AUROC of 0.94 [0.91-0.98] with 94% (75/80) sensitivity and 77% (92/120) specificity at PI-RADS >= 3. The deep radiomics model at a tumor probability cut-off >= 0.76 achieved 0.91 [0.86-0.95] AUROC with 90% (72/80) sensitivity and 73% (87/120) specificity, not significantly different (p = 0.068) from PI-RADS. On the lesion level, PI-RADS cut-off >= 3 had 84% (91/108) sensitivity at 0.2 (40/200) false positives per patient, while deep radiomics attained 68% (73/108) sensitivity at the same false positive rate. Conclusion: Deep radiomics machine learning model achieved comparable performance to PI-RADS assessment in csPCa detection at the patient-level but not at the lesion-level.

comment: 20 pages, 4 figures, 4 tables

☆ LLaVA-KD: A Framework of Distilling Multimodal Large Language Models

Yuxuan Cai, Jiangning Zhang, Haoyang He, Xinwei He, Ao Tong, Zhenye Gan, Chengjie Wang, Xiang Bai

The success of Large Language Models (LLM) has led researchers to explore Multimodal Large Language Models (MLLM) for unified visual and linguistic understanding. However, the increasing model size and computational complexity of MLLM limit their use in resource-constrained environments. Small-scale MLLM (s-MLLM) aims to retain the capabilities of the large-scale model (l-MLLM) while reducing computational demands, but resulting in a significant decline in performance. To address the aforementioned issues, we propose a novel LLaVA-KD framework to transfer knowledge from l-MLLM to s-MLLM. Specifically, we introduce Multimodal Distillation (MDist) to minimize the divergence between the visual-textual output distributions of l-MLLM and s-MLLM, and Relation Distillation (RDist) to transfer l-MLLM's ability to model correlations between visual features. Additionally, we propose a three-stage training scheme to fully exploit the potential of s-MLLM: 1) Distilled Pre-Training to align visual-textual representations, 2) Supervised Fine-Tuning to equip the model with multimodal understanding, and 3) Distilled Fine-Tuning to further transfer l-MLLM capabilities. Our approach significantly improves performance without altering the small model's architecture. Extensive experiments and ablation studies validate the effectiveness of each proposed component. Code will be available at https://github.com/caiyuxuan1120/LLaVA-KD.

comment: Under review

☆ Managing Bandwidth: The Key to Cloud-Assisted Autonomous Driving

Alexander Krentsel, Peter Schafhalter, Joseph E. Gonzalez, Sylvia Ratnasamy, Scott Shenker, Ion Stoica

Prevailing wisdom asserts that one cannot rely on the cloud for critical real-time control systems like self-driving cars. We argue that we can, and must. Following the trends of increasing model sizes, improvements in hardware, and evolving mobile networks, we identify an opportunity to offload parts of time-sensitive and latency-critical compute to the cloud. Doing so requires carefully allocating bandwidth to meet strict latency SLOs, while maximizing benefit to the car.

comment: 6 pages

☆ Improve Vision Language Model Chain-of-thought Reasoning

Ruohong Zhang, Bowen Zhang, Yanghao Li, Haotian Zhang, Zhiqing Sun, Zhe Gan, Yinfei Yang, Ruoming Pang, Yiming Yang

Chain-of-thought (CoT) reasoning in vision language models (VLMs) is crucial for improving interpretability and trustworthiness. However, current training recipes lack robust CoT reasoning data, relying on datasets dominated by short annotations with minimal rationales. In this work, we show that training VLM on short answers does not generalize well to reasoning tasks that require more detailed responses. To address this, we propose a two-fold approach. First, we distill rationales from GPT-4o model to enrich the training data and fine-tune VLMs, boosting their CoT performance. Second, we apply reinforcement learning to further calibrate reasoning quality. Specifically, we construct positive (correct) and negative (incorrect) pairs of model-generated reasoning chains, by comparing their predictions with annotated short answers. Using this pairwise data, we apply the Direct Preference Optimization algorithm to refine the model's reasoning abilities. Our experiments demonstrate significant improvements in CoT reasoning on benchmark datasets and better generalization to direct answer prediction as well. This work emphasizes the importance of incorporating detailed rationales in training and leveraging reinforcement learning to strengthen the reasoning capabilities of VLMs.

comment: 10 pages + appendix

☆ Training Better Deep Learning Models Using Human Saliency

Aidan Boyd, Patrick Tinsley, Kevin W. Bowyer, Adam Czajka

This work explores how human judgement about salient regions of an image can be introduced into deep convolutional neural network (DCNN) training. Traditionally, training of DCNNs is purely data-driven. This often results in learning features of the data that are only coincidentally correlated with class labels. Human saliency can guide network training using our proposed new component of the loss function that ConveYs Brain Oversight to Raise Generalization (CYBORG) and penalizes the model for using non-salient regions. This mechanism produces DCNNs achieving higher accuracy and generalization compared to using the same training data without human salience. Experimental results demonstrate that CYBORG applies across multiple network architectures and problem domains (detection of synthetic faces, iris presentation attacks and anomalies in chest X-rays), while requiring significantly less data than training without human saliency guidance. Visualizations show that CYBORG-trained models' saliency is more consistent across independent training runs than traditionally-trained models, and also in better agreement with humans. To lower the cost of collecting human annotations, we also explore using deep learning to provide automated annotations. CYBORG training of CNNs addresses important issues such as reducing the appetite for large training sets, increasing interpretability, and reducing fragility by generalizing better to new types of data.

☆ A Framework for Evaluating Predictive Models Using Synthetic Image Covariates and Longitudinal Data

Simon Deltadahl, Andreu Vall, Vijay Ivaturi, Niklas Korsbo

We present a novel framework for synthesizing patient data with complex covariates (e.g., eye scans) paired with longitudinal observations (e.g., visual acuity over time), addressing privacy concerns in healthcare research. Our approach introduces controlled association in latent spaces generating each data modality, enabling the creation of complex covariate-longitudinal observation pairs. This framework facilitates the development of predictive models and provides openly available benchmarking datasets for healthcare research. We demonstrate our framework using optical coherence tomography (OCT) scans, though it is applicable across domains. Using 109,309 2D OCT scan slices, we trained an image generative model combining a variational autoencoder and a diffusion model. Longitudinal observations were simulated using a nonlinear mixed effect (NLME) model from a low-dimensional space of random effects. We generated 1.1M OCT scan slices paired with five sets of longitudinal observations at controlled association levels (100%, 50%, 10%, 5.26%, and 2% of between-subject variability). To assess the framework, we modeled synthetic longitudinal observations with another NLME model, computed empirical Bayes estimates of random effects, and trained a ResNet to predict these estimates from synthetic OCT scans. We then incorporated ResNet predictions into the NLME model for patient-individualized predictions. Prediction accuracy on withheld data declined as intended with reduced association between images and longitudinal measurements. Notably, in all but the 2% case, we achieved within 50% of the theoretical best possible prediction on withheld data, demonstrating our ability to detect even weak signals. This confirms the effectiveness of our framework in generating synthetic data with controlled levels of association, providing a valuable tool for healthcare research.

☆ Beyond Filtering: Adaptive Image-Text Quality Enhancement for MLLM Pretraining

Han Huang, Yuqi Huo, Zijia Zhao, Haoyu Lu, Shu Wu, Bingning Wang, Qiang Liu, Weipeng Chen, Liang Wang

Multimodal large language models (MLLMs) have made significant strides by integrating visual and textual modalities. A critical factor in training MLLMs is the quality of image-text pairs within multimodal pretraining datasets. However, $\textit {de facto}$ filter-based data quality enhancement paradigms often discard a substantial portion of high-quality image data due to inadequate semantic alignment between images and texts, leading to inefficiencies in data utilization and scalability. In this paper, we propose the Adaptive Image-Text Quality Enhancer (AITQE), a model that dynamically assesses and enhances the quality of image-text pairs. AITQE employs a text rewriting mechanism for low-quality pairs and incorporates a negative sample learning strategy to improve evaluative capabilities by integrating deliberately selected low-quality samples during training. Unlike prior approaches that significantly alter text distributions, our method minimally adjusts text to preserve data volume while enhancing quality. Experimental results demonstrate that AITQE surpasses existing methods on various benchmark, effectively leveraging raw data and scaling efficiently with increasing data volumes. We hope our work will inspire future works. The code and model are available at: https://github.com/hanhuang22/AITQE.

☆ Griffon-G: Bridging Vision-Language and Vision-Centric Tasks via Large Multimodal Models

Yufei Zhan, Hongyin Zhao, Yousong Zhu, Fan Yang, Ming Tang, Jinqiao Wang

Large Multimodal Models (LMMs) have achieved significant breakthroughs in various vision-language and vision-centric tasks based on auto-regressive modeling. However, these models typically focus on either vision-centric tasks, such as visual grounding and region description, or vision-language tasks, like image caption and multi-scenario VQAs. None of the LMMs have yet comprehensively unified both types of tasks within a single model, as seen in Large Language Models in the natural language processing field. Furthermore, even with abundant multi-task instruction-following data, directly stacking these data for universal capabilities extension remains challenging. To address these issues, we introduce a novel multi-dimension curated and consolidated multimodal dataset, named CCMD-8M, which overcomes the data barriers of unifying vision-centric and vision-language tasks through multi-level data curation and multi-task consolidation. More importantly, we present Griffon-G, a general large multimodal model that addresses both vision-centric and vision-language tasks within a single end-to-end paradigm. Griffon-G resolves the training collapse issue encountered during the joint optimization of these tasks, achieving better training efficiency. Evaluations across multimodal benchmarks, general Visual Question Answering (VQA) tasks, scene text-centric VQA tasks, document-related VQA tasks, Referring Expression Comprehension, and object detection demonstrate that Griffon-G surpasses the advanced LMMs and achieves expert-level performance in complicated vision-centric tasks.

comment: This work has been submitted to the IEEE for possible publication. Codes and data will be later released at https://github.com/jefferyZhan/Griffon

☆ Sparkle: Mastering Basic Spatial Capabilities in Vision Language Models Elicits Generalization to Composite Spatial Reasoning

Yihong Tang, Ao Qu, Zhaokai Wang, Dingyi Zhuang, Zhaofeng Wu, Wei Ma, Shenhao Wang, Yunhan Zheng, Zhan Zhao, Jinhua Zhao

Vision language models (VLMs) have demonstrated impressive performance across a wide range of downstream tasks. However, their proficiency in spatial reasoning remains limited, despite its crucial role in tasks involving navigation and interaction with physical environments. Specifically, much of the spatial reasoning in these tasks occurs in two-dimensional (2D) environments, and our evaluation reveals that state-of-the-art VLMs frequently generate implausible and incorrect responses to composite spatial reasoning problems, including simple pathfinding tasks that humans can solve effortlessly at a glance. To address this, we explore an effective approach to enhance 2D spatial reasoning within VLMs by training the model on basic spatial capabilities. We begin by disentangling the key components of 2D spatial reasoning: direction comprehension, distance estimation, and localization. Our central hypothesis is that mastering these basic spatial capabilities can significantly enhance a model's performance on composite spatial tasks requiring advanced spatial understanding and combinatorial problem-solving. To investigate this hypothesis, we introduce Sparkle, a framework that fine-tunes VLMs on these three basic spatial capabilities by synthetic data generation and targeted supervision to form an instruction dataset for each capability. Our experiments demonstrate that VLMs fine-tuned with Sparkle achieve significant performance gains, not only in the basic tasks themselves but also in generalizing to composite and out-of-distribution spatial reasoning tasks (e.g., improving from 13.5% to 40.0% on the shortest path problem). These findings underscore the effectiveness of mastering basic spatial capabilities in enhancing composite spatial problem-solving, offering insights for improving VLMs' spatial reasoning capabilities.

☆ Metric as Transform: Exploring beyond Affine Transform for Interpretable Neural Network

Suman Sapkota

Artificial Neural Networks of varying architectures are generally paired with affine transformation at the core. However, we find dot product neurons with global influence less interpretable as compared to local influence of euclidean distance (as used in Radial Basis Function Network). In this work, we explore the generalization of dot product neurons to $l^p$-norm, metrics, and beyond. We find that metrics as transform performs similarly to affine transform when used in MultiLayer Perceptron or Convolutional Neural Network. Moreover, we explore various properties of Metrics, compare it with Affine, and present multiple cases where metrics seem to provide better interpretability. We develop an interpretable local dictionary based Neural Networks and use it to understand and reject adversarial examples.

comment: 22 pages, 20 figures, 3 tables

☆ Pangea: A Fully Open Multilingual Multimodal LLM for 39 Languages

Xiang Yue, Yueqi Song, Akari Asai, Seungone Kim, Jean de Dieu Nyandwi, Simran Khanuja, Anjali Kantharuban, Lintang Sutawika, Sathyanarayanan Ramamoorthy, Graham Neubig

Despite recent advances in multimodal large language models (MLLMs), their development has predominantly focused on English- and western-centric datasets and tasks, leaving most of the world's languages and diverse cultural contexts underrepresented. This paper introduces Pangea, a multilingual multimodal LLM trained on PangeaIns, a diverse 6M instruction dataset spanning 39 languages. PangeaIns features: 1) high-quality English instructions, 2) carefully machine-translated instructions, and 3) culturally relevant multimodal tasks to ensure cross-cultural coverage. To rigorously assess models' capabilities, we introduce PangeaBench, a holistic evaluation suite encompassing 14 datasets covering 47 languages. Results show that Pangea significantly outperforms existing open-source models in multilingual settings and diverse cultural contexts. Ablation studies further reveal the importance of English data proportions, language popularity, and the number of multimodal training samples on overall performance. We fully open-source our data, code, and trained checkpoints, to facilitate the development of inclusive and robust multilingual MLLMs, promoting equity and accessibility across a broader linguistic and cultural spectrum.

comment: 52 pages, 27 figures

☆ Warped Diffusion: Solving Video Inverse Problems with Image Diffusion Models NeurIPS 2024

Giannis Daras, Weili Nie, Karsten Kreis, Alex Dimakis, Morteza Mardani, Nikola Borislavov Kovachki, Arash Vahdat

Using image models naively for solving inverse video problems often suffers from flickering, texture-sticking, and temporal inconsistency in generated videos. To tackle these problems, in this paper, we view frames as continuous functions in the 2D space, and videos as a sequence of continuous warping transformations between different frames. This perspective allows us to train function space diffusion models only on images and utilize them to solve temporally correlated inverse problems. The function space diffusion models need to be equivariant with respect to the underlying spatial transformations. To ensure temporal consistency, we introduce a simple post-hoc test-time guidance towards (self)-equivariant solutions. Our method allows us to deploy state-of-the-art latent diffusion models such as Stable Diffusion XL to solve video inverse problems. We demonstrate the effectiveness of our method for video inpainting and $8\times$ video super-resolution, outperforming existing techniques based on noise transformations. We provide generated video results: https://giannisdaras.github.io/warped\_diffusion.github.io/.

comment: Accepted in NeurIPS 2024

☆ Towards Combating Frequency Simplicity-biased Learning for Domain Generalization NeurIPS 2024

Xilin He, Jingyu Hu, Qinliang Lin, Cheng Luo, Weicheng Xie, Siyang Song, Muhammad Haris Khan, Linlin Shen

Domain generalization methods aim to learn transferable knowledge from source domains that can generalize well to unseen target domains. Recent studies show that neural networks frequently suffer from a simplicity-biased learning behavior which leads to over-reliance on specific frequency sets, namely as frequency shortcuts, instead of semantic information, resulting in poor generalization performance. Despite previous data augmentation techniques successfully enhancing generalization performances, they intend to apply more frequency shortcuts, thereby causing hallucinations of generalization improvement. In this paper, we aim to prevent such learning behavior of applying frequency shortcuts from a data-driven perspective. Given the theoretical justification of models' biased learning behavior on different spatial frequency components, which is based on the dataset frequency properties, we argue that the learning behavior on various frequency components could be manipulated by changing the dataset statistical structure in the Fourier domain. Intuitively, as frequency shortcuts are hidden in the dominant and highly dependent frequencies of dataset structure, dynamically perturbating the over-reliance frequency components could prevent the application of frequency shortcuts. To this end, we propose two effective data augmentation modules designed to collaboratively and adaptively adjust the frequency characteristic of the dataset, aiming to dynamically influence the learning behavior of the model and ultimately serving as a strategy to mitigate shortcut learning. Code is available at AdvFrequency (https://github.com/C0notSilly/AdvFrequency).

comment: Accepted by NeurIPS 2024

☆ An Explainable Contrastive-based Dilated Convolutional Network with Transformer for Pediatric Pneumonia Detection

Chandravardhan Singh Raghaw, Parth Shirish Bhore, Mohammad Zia Ur Rehman, Nagendra Kumar

Pediatric pneumonia remains a significant global threat, posing a larger mortality risk than any other communicable disease. According to UNICEF, it is a leading cause of mortality in children under five and requires prompt diagnosis. Early diagnosis using chest radiographs is the prevalent standard, but limitations include low radiation levels in unprocessed images and data imbalance issues. This necessitates the development of efficient, computer-aided diagnosis techniques. To this end, we propose a novel EXplainable Contrastive-based Dilated Convolutional Network with Transformer (XCCNet) for pediatric pneumonia detection. XCCNet harnesses the spatial power of dilated convolutions and the global insights from contrastive-based transformers for effective feature refinement. A robust chest X-ray processing module tackles low-intensity radiographs, while adversarial-based data augmentation mitigates the skewed distribution of chest X-rays in the dataset. Furthermore, we actively integrate an explainability approach through feature visualization, directly aligning it with the attention region that pinpoints the presence of pneumonia or normality in radiographs. The efficacy of XCCNet is comprehensively assessed on four publicly available datasets. Extensive performance evaluation demonstrates the superiority of XCCNet compared to state-of-the-art methods.

☆ Multimodal Flare Forecasting with Deep Learning

Grégoire Francisco, Sabrina Guastavino, Teresa Barata, João Fernandes, Dario Del Moro

Solar flare forecasting mainly relies on photospheric magnetograms and associated physical features to predict forthcoming flares. However, it is believed that flare initiation mechanisms often originate in the chromosphere and the lower corona. In this study, we employ deep learning as a purely data-driven approach to compare the predictive capabilities of chromospheric and coronal UV and EUV emissions across different wavelengths with those of photospheric line-of-sight magnetograms. Our findings indicate that individual EUV wavelengths can provide discriminatory power comparable or better to that of line-of-sight magnetograms. Moreover, we identify simple multimodal neural network architectures that consistently outperform single-input models, showing complementarity between the flare precursors that can be extracted from the distinct layers of the solar atmosphere. To mitigate potential biases from known misattributions in Active Region flare catalogs, our models are trained and evaluated using full-disk images and a comprehensive flare event catalog at the full-disk level. We introduce a deep-learning architecture suited for extracting temporal features from full-disk videos.

☆ Increasing Interpretability of Neural Networks By Approximating Human Visual Saliency

Aidan Boyd, Mohamed Trabelsi, Huseyin Uzunalioglu, Dan Kushnir

Understanding specifically where a model focuses on within an image is critical for human interpretability of the decision-making process. Deep learning-based solutions are prone to learning coincidental correlations in training datasets, causing over-fitting and reducing the explainability. Recent advances have shown that guiding models to human-defined regions of saliency within individual images significantly increases performance and interpretability. Human-guided models also exhibit greater generalization capabilities, as coincidental dataset features are avoided. Results show that models trained with saliency incorporation display an increase in interpretability of up to 30% over models trained without saliency information. The collection of this saliency information, however, can be costly, laborious and in some cases infeasible. To address this limitation, we propose a combination strategy of saliency incorporation and active learning to reduce the human annotation data required by 80% while maintaining the interpretability and performance increase from human saliency. Extensive experimentation outlines the effectiveness of the proposed approach across five public datasets and six active learning criteria.

☆ LMHaze: Intensity-aware Image Dehazing with a Large-scale Multi-intensity Real Haze Dataset

Ruikun Zhang, Hao Yang, Yan Yang, Ying Fu, Liyuan Pan

Image dehazing has drawn a significant attention in recent years. Learning-based methods usually require paired hazy and corresponding ground truth (haze-free) images for training. However, it is difficult to collect real-world image pairs, which prevents developments of existing methods. Although several works partially alleviate this issue by using synthetic datasets or small-scale real datasets. The haze intensity distribution bias and scene homogeneity in existing datasets limit the generalization ability of these methods, particularly when encountering images with previously unseen haze intensities. In this work, we present LMHaze, a large-scale, high-quality real-world dataset. LMHaze comprises paired hazy and haze-free images captured in diverse indoor and outdoor environments, spanning multiple scenarios and haze intensities. It contains over 5K high-resolution image pairs, surpassing the size of the biggest existing real-world dehazing dataset by over 25 times. Meanwhile, to better handle images with different haze intensities, we propose a mixture-of-experts model based on Mamba (MoE-Mamba) for dehazing, which dynamically adjusts the model parameters according to the haze intensity. Moreover, with our proposed dataset, we conduct a new large multimodal model (LMM)-based benchmark study to simulate human perception for evaluating dehazed images. Experiments demonstrate that LMHaze dataset improves the dehazing performance in real scenarios and our dehazing method provides better results compared to state-of-the-art methods.

☆ Final Report for CHESS: Cloud, High-Performance Computing, and Edge for Science and Security

Nathan Tallent, Jan Strube, Luanzheng Guo, Hyungro Lee, Jesun Firoz, Sayan Ghosh, Bo Fang, Oceane Bel, Steven Spurgeon, Sarah Akers, Christina Doty, Erol Cromwell

Automating the theory-experiment cycle requires effective distributed workflows that utilize a computing continuum spanning lab instruments, edge sensors, computing resources at multiple facilities, data sets distributed across multiple information sources, and potentially cloud. Unfortunately, the obvious methods for constructing continuum platforms, orchestrating workflow tasks, and curating datasets over time fail to achieve scientific requirements for performance, energy, security, and reliability. Furthermore, achieving the best use of continuum resources depends upon the efficient composition and execution of workflow tasks, i.e., combinations of numerical solvers, data analytics, and machine learning. Pacific Northwest National Laboratory's LDRD "Cloud, High-Performance Computing (HPC), and Edge for Science and Security" (CHESS) has developed a set of interrelated capabilities for enabling distributed scientific workflows and curating datasets. This report describes the results and successes of CHESS from the perspective of open science.

☆ Integrated Image-Text Based on Semi-supervised Learning for Small Sample Instance Segmentation

Ruting Chi, Zhiyi Huang, Yuexing Han

Small sample instance segmentation is a very challenging task, and many existing methods follow the training strategy of meta-learning which pre-train models on support set and fine-tune on query set. The pre-training phase, which is highly task related, requires a significant amount of additional training time and the selection of datasets with close proximity to ensure effectiveness. The article proposes a novel small sample instance segmentation solution from the perspective of maximizing the utilization of existing information without increasing annotation burden and training costs. The proposed method designs two modules to address the problems encountered in small sample instance segmentation. First, it helps the model fully utilize unlabeled data by learning to generate pseudo labels, increasing the number of available samples. Second, by integrating the features of text and image, more accurate classification results can be obtained. These two modules are suitable for box-free and box-dependent frameworks. In the way, the proposed method not only improves the performance of small sample instance segmentation, but also greatly reduce reliance on pre-training. We have conducted experiments in three datasets from different scenes: on land, underwater and under microscope. As evidenced by our experiments, integrated image-text corrects the confidence of classification, and pseudo labels help the model obtain preciser masks. All the results demonstrate the effectiveness and superiority of our method.

☆ Label Filling via Mixed Supervision for Medical Image Segmentation from Noisy Annotations

Ming Li, Wei Shen, Qingli Li, Yan Wang

The success of medical image segmentation usually requires a large number of high-quality labels. But since the labeling process is usually affected by the raters' varying skill levels and characteristics, the estimated masks provided by different raters usually suffer from high inter-rater variability. In this paper, we propose a simple yet effective Label Filling framework, termed as LF-Net, predicting the groundtruth segmentation label given only noisy annotations during training. The fundamental idea of label filling is to supervise the segmentation model by a subset of pixels with trustworthy labels, meanwhile filling labels of other pixels by mixed supervision. More concretely, we propose a qualified majority voting strategy, i.e., a threshold voting scheme is designed to model agreement among raters and the majority-voted labels of the selected subset of pixels are regarded as supervision. To fill labels of other pixels, two types of mixed auxiliary supervision are proposed: a soft label learned from intrinsic structures of noisy annotations, and raters' characteristics labels which propagate individual rater's characteristics information. LF-Net has two main advantages. 1) Training with trustworthy pixels incorporates training with confident supervision, guiding the direction of groundtruth label learning. 2) Two types of mixed supervision prevent over-fitting issues when the network is supervised by a subset of pixels, and guarantee high fidelity with the true label. Results on five datasets of diverse imaging modalities show that our LF-Net boosts segmentation accuracy in all datasets compared with state-of-the-art methods, with even a 7% improvement in DSC for MS lesion segmentation.

☆ Benchmarking Pathology Foundation Models: Adaptation Strategies and Scenarios

Jeaung Lee, Jeewoo Lim, Keunho Byeon, Jin Tae Kwak

In computational pathology, several foundation models have recently emerged and demonstrated enhanced learning capability for analyzing pathology images. However, adapting these models to various downstream tasks remains challenging, particularly when faced with datasets from different sources and acquisition conditions, as well as limited data availability. In this study, we benchmark four pathology-specific foundation models across 14 datasets and two scenarios-consistency assessment and flexibility assessment-addressing diverse adaptation scenarios and downstream tasks. In the consistency assessment scenario, involving five fine-tuning methods, we found that the parameter-efficient fine-tuning approach was both efficient and effective for adapting pathology-specific foundation models to diverse datasets within the same downstream task. In the flexibility assessment scenario under data-limited environments, utilizing five few-shot learning methods, we observed that the foundation models benefited more from the few-shot learning methods that involve modification during the testing phase only. These findings provide insights that could guide the deployment of pathology-specific foundation models in real clinical settings, potentially improving the accuracy and reliability of pathology image analysis. The code for this study is available at: https://github.com/QuIIL/BenchmarkingPathologyFoundationModels.

☆ Improving the Multi-label Atomic Activity Recognition by Robust Visual Feature and Advanced Attention @ ROAD++ Atomic Activity Recognition 2024

Jiamin Cao, Lingqi Wang, Kexin Zhang, Yuting Yang, Licheng Jiao, Yuwei Guo

Road++ Track3 proposes a multi-label atomic activity recognition task in traffic scenarios, which can be standardized as a 64-class multi-label video action recognition task. In the multi-label atomic activity recognition task, the robustness of visual feature extraction remains a key challenge, which directly affects the model performance and generalization ability. To cope with these issues, our team optimized three aspects: data processing, model and post-processing. Firstly, the appropriate resolution and video sampling strategy are selected, and a fixed sampling strategy is set on the validation and test sets. Secondly, in terms of model training, the team selects a variety of visual backbone networks for feature extraction, and then introduces the action-slot model, which is trained on the training and validation sets, and reasoned on the test set. Finally, for post-processing, the team combined the strengths and weaknesses of different models for weighted fusion, and the final mAP on the test set was 58%, which is 4% higher than the challenge baseline.

☆ Few-shot target-driven instance detection based on open-vocabulary object detection models

Ben Crulis, Barthelemy Serres, Cyril De Runz, Gilles Venturini

Current large open vision models could be useful for one and few-shot object recognition. Nevertheless, gradient-based re-training solutions are costly. On the other hand, open-vocabulary object detection models bring closer visual and textual concepts in the same latent space, allowing zero-shot detection via prompting at small computational cost. We propose a lightweight method to turn the latter into a one-shot or few-shot object recognition models without requiring textual descriptions. Our experiments on the TEgO dataset using the YOLO-World model as a base show that performance increases with the model size, the number of examples and the use of image augmentation.

☆ START: A Generalized State Space Model with Saliency-Driven Token-Aware Transformation NeurIPS2024

Jintao Guo, Lei Qi, Yinghuan Shi, Yang Gao

Domain Generalization (DG) aims to enable models to generalize to unseen target domains by learning from multiple source domains. Existing DG methods primarily rely on convolutional neural networks (CNNs), which inherently learn texture biases due to their limited receptive fields, making them prone to overfitting source domains. While some works have introduced transformer-based methods (ViTs) for DG to leverage the global receptive field, these methods incur high computational costs due to the quadratic complexity of self-attention. Recently, advanced state space models (SSMs), represented by Mamba, have shown promising results in supervised learning tasks by achieving linear complexity in sequence length during training and fast RNN-like computation during inference. Inspired by this, we investigate the generalization ability of the Mamba model under domain shifts and find that input-dependent matrices within SSMs could accumulate and amplify domain-specific features, thus hindering model generalization. To address this issue, we propose a novel SSM-based architecture with saliency-based token-aware transformation (namely START), which achieves state-of-the-art (SOTA) performances and offers a competitive alternative to CNNs and ViTs. Our START can selectively perturb and suppress domain-specific features in salient tokens within the input-dependent matrices of SSMs, thus effectively reducing the discrepancy between different domains. Extensive experiments on five benchmarks demonstrate that START outperforms existing SOTA DG methods with efficient linear complexity. Our code is available at https://github.com/lingeringlight/START.

comment: Accepted by NeurIPS2024. The code is available at https://github.com/lingeringlight/START

☆ Multispectral Texture Synthesis using RGB Convolutional Neural Networks

Sélim Ollivier, Yann Gousseau, Sidonie Lefebvre

State-of-the-art RGB texture synthesis algorithms rely on style distances that are computed through statistics of deep features. These deep features are extracted by classification neural networks that have been trained on large datasets of RGB images. Extending such synthesis methods to multispectral images is not straightforward, since the pre-trained networks are designed for and have been trained on RGB images. In this work, we propose two solutions to extend these methods to multispectral imaging. Neither of them require additional training of the neural network from which the second order neural statistics are extracted. The first one consists in optimizing over batches of random triplets of spectral bands throughout training. The second one projects multispectral pixels onto a 3 dimensional space. We further explore the benefit of a color transfer operation upstream of the projection to avoid the potentially abnormal color distributions induced by the projection. Our experiments compare the performances of the various methods through different metrics. We demonstrate that they can be used to perform exemplar-based texture synthesis, achieve good visual quality and comes close to state-of-the art methods on RGB bands.

☆ Massimo: Public Queue Monitoring and Management using Mass-Spring Model

Abhijeet Kumar, Unnati Singh, Rajdeep Chatterjee, Tathagata Bandyopadhyay

An efficient system of a queue control and regulation in public spaces is very important in order to avoid the traffic jams and to improve the customer satisfaction. This article offers a detailed road map based on a merger of intelligent systems and creating an efficient systems of queues in public places. Through the utilization of different technologies i.e. computer vision, machine learning algorithms, deep learning our system provide accurate information about the place is crowded or not and the necessary efforts to be taken.

comment: 8 pages, 6 figures, 3 algorithms, 3 tables

☆ 3D-GANTex: 3D Face Reconstruction with StyleGAN3-based Multi-View Images and 3DDFA based Mesh Generation

Rohit Das, Tzung-Han Lin, Ko-Chih Wang

Geometry and texture estimation from a single face image is an ill-posed problem since there is very little information to work with. The problem further escalates when the face is rotated at a different angle. This paper tries to tackle this problem by introducing a novel method for texture estimation from a single image by first using StyleGAN and 3D Morphable Models. The method begins by generating multi-view faces using the latent space of GAN. Then 3DDFA trained on 3DMM estimates a 3D face mesh as well as a high-resolution texture map that is consistent with the estimated face shape. The result shows that the generated mesh is of high quality with near to accurate texture representation.

comment: 7 pages, 4 figures, 2 tables, pre-print version

☆ Visual Representation Learning Guided By Multi-modal Prior Knowledge

Hongkuan Zhou, Lavdim Halilaj, Sebastian Monka, Stefan Schmid, Yuqicheng Zhu, Bo Xiong, Steffen Staab

Despite the remarkable success of deep neural networks (DNNs) in computer vision, they fail to remain high-performing when facing distribution shifts between training and testing data. In this paper, we propose Knowledge-Guided Visual representation learning (KGV), a distribution-based learning approach leveraging multi-modal prior knowledge, to improve generalization under distribution shift. We use prior knowledge from two distinct modalities: 1) a knowledge graph (KG) with hierarchical and association relationships; and 2) generated synthetic images of visual elements semantically represented in the KG. The respective embeddings are generated from the given modalities in a common latent space, i.e., visual embeddings from original and synthetic images as well as knowledge graph embeddings (KGEs). These embeddings are aligned via a novel variant of translation-based KGE methods, where the node and relation embeddings of the KG are modeled as Gaussian distributions and translations respectively. We claim that incorporating multi-model prior knowledge enables more regularized learning of image representations. Thus, the models are able to better generalize across different data distributions. We evaluate KGV on different image classification tasks with major or minor distribution shifts, namely road sign classification across datasets from Germany, China, and Russia, image classification with the mini-ImageNet dataset and its variants, as well as the DVM-CAR dataset. The results demonstrate that KGV consistently exhibits higher accuracy and data efficiency than the baselines across all experiments.

☆ Granularity Matters in Long-Tail Learning

Shizhen Zhao, Xin Wen, Jiahui Liu, Chuofan Ma, Chunfeng Yuan, Xiaojuan Qi

Balancing training on long-tail data distributions remains a long-standing challenge in deep learning. While methods such as re-weighting and re-sampling help alleviate the imbalance issue, limited sample diversity continues to hinder models from learning robust and generalizable feature representations, particularly for tail classes. In contrast to existing methods, we offer a novel perspective on long-tail learning, inspired by an observation: datasets with finer granularity tend to be less affected by data imbalance. In this paper, we investigate this phenomenon through both quantitative and qualitative studies, showing that increased granularity enhances the generalization of learned features in tail categories. Motivated by these findings, we propose a method to increase dataset granularity through category extrapolation. Specifically, we introduce open-set auxiliary classes that are visually similar to existing ones, aiming to enhance representation learning for both head and tail classes. This forms the core contribution and insight of our approach. To automate the curation of auxiliary data, we leverage large language models (LLMs) as knowledge bases to search for auxiliary categories and retrieve relevant images through web crawling. To prevent the overwhelming presence of auxiliary classes from disrupting training, we introduce a neighbor-silencing loss that encourages the model to focus on class discrimination within the target dataset. During inference, the classifier weights for auxiliary categories are masked out, leaving only the target class weights for use. Extensive experiments and ablation studies on three standard long-tail benchmarks demonstrate the effectiveness of our approach, notably outperforming strong baseline methods that use the same amount of data. The code will be made publicly available.

☆ Zero-Shot Scene Reconstruction from Single Images with Deep Prior Assembly NeurIPS 2024

Junsheng Zhou, Yu-Shen Liu, Zhizhong Han

Large language and vision models have been leading a revolution in visual computing. By greatly scaling up sizes of data and model parameters, the large models learn deep priors which lead to remarkable performance in various tasks. In this work, we present deep prior assembly, a novel framework that assembles diverse deep priors from large models for scene reconstruction from single images in a zero-shot manner. We show that this challenging task can be done without extra knowledge but just simply generalizing one deep prior in one sub-task. To this end, we introduce novel methods related to poses, scales, and occlusion parsing which are keys to enable deep priors to work together in a robust way. Deep prior assembly does not require any 3D or 2D data-driven training in the task and demonstrates superior performance in generalizing priors to open-world scenes. We conduct evaluations on various datasets, and report analysis, numerical and visual comparisons with the latest methods to show our superiority. Project page: https://junshengzhou.github.io/DeepPriorAssembly.

comment: To appear at NeurIPS 2024. Project page: https://junshengzhou.github.io/DeepPriorAssembly

☆ A Paradigm Shift in Mouza Map Vectorization: A Human-Machine Collaboration Approach

Mahir Shahriar Dhrubo, Samira Akter, Anwarul Bashir Shuaib, Md Toki Tahmid, Zahid Hasan, A. B. M. Alim Al Islam

Efficient vectorization of hand-drawn cadastral maps, such as Mouza maps in Bangladesh, poses a significant challenge due to their complex structures. Current manual digitization methods are time-consuming and labor-intensive. Our study proposes a semi-automated approach to streamline the digitization process, saving both time and human resources. Our methodology focuses on separating the plot boundaries and plot identifiers and applying our digitization methodology to convert both of them into vectorized format. To accomplish full vectorization, Convolutional Neural Network (CNN) models are utilized for pre-processing and plot number detection along with our smoothing algorithms based on the diversity of vector maps. The CNN models are trained with our own labeled dataset, generated from the maps, and smoothing algorithms are introduced from the various observations of the map's vector formats. Further human intervention remains essential for precision. We have evaluated our methods on several maps and provided both quantitative and qualitative results with user study. The result demonstrates that our methodology outperforms the existing map digitization processes significantly.

comment: 13 pages including reference, 14 figures, 4 tables

☆ Diffusion Transformer Policy

Zhi Hou, Tianyi Zhang, Yuwen Xiong, Hengjun Pu, Chengyang Zhao, Ronglei Tong, Yu Qiao, Jifeng Dai, Yuntao Chen

Recent large visual-language action models pretrained on diverse robot datasets have demonstrated the potential for generalizing to new environments with a few in-domain data. However, those approaches usually predict discretized or continuous actions by a small action head, which limits the ability in handling diverse action spaces. In contrast, we model the continuous action with a large multi-modal diffusion transformer, dubbed as Diffusion Transformer Policy, in which we directly denoise action chunks by a large transformer model rather than a small action head. By leveraging the scaling capability of transformers, the proposed approach can effectively model continuous end-effector actions across large diverse robot datasets, and achieve better generalization performance. Extensive experiments demonstrate Diffusion Transformer Policy pretrained on diverse robot data can generalize to different embodiments, including simulation environments like Maniskill2 and Calvin, as well as the real-world Franka arm. Specifically, without bells and whistles, the proposed approach achieves state-of-the-art performance with only a single third-view camera stream in the Calvin novel task setting (ABC->D), improving the average number of tasks completed in a row of 5 to 3.6, and the pretraining stage significantly facilitates the success sequence length on the Calvin by over 1.2. The code will be publicly available.

comment: Preprint

☆ CamI2V: Camera-Controlled Image-to-Video Diffusion Model

Guangcong Zheng, Teng Li, Rui Jiang, Yehao Lu, Tao Wu, Xi Li

Recently, camera pose, as a user-friendly and physics-related condition, has been introduced into text-to-video diffusion model for camera control. However, existing methods simply inject camera conditions through a side input. These approaches neglect the inherent physical knowledge of camera pose, resulting in imprecise camera control, inconsistencies, and also poor interpretability. In this paper, we emphasize the necessity of integrating explicit physical constraints into model design. Epipolar attention is proposed for modeling all cross-frame relationships from a novel perspective of noised condition. This ensures that features are aggregated from corresponding epipolar lines in all noised frames, overcoming the limitations of current attention mechanisms in tracking displaced features across frames, especially when features move significantly with the camera and become obscured by noise. Additionally, we introduce register tokens to handle cases without intersections between frames, commonly caused by rapid camera movements, dynamic objects, or occlusions. To support image-to-video, we propose the multiple guidance scale to allow for precise control for image, text, and camera, respectively. Furthermore, we establish a more robust and reproducible evaluation pipeline to solve the inaccuracy and instability of existing camera control measurement. We achieve a 25.5\% improvement in camera controllability on RealEstate10K while maintaining strong generalization to out-of-domain images. Only 24GB and 12GB are required for training and inference, respectively. We plan to release checkpoints, along with training and evaluation codes. Dynamic videos are best viewed at \url{https://zgctroy.github.io/CamI2V}.

☆ AI-Driven Approaches for Glaucoma Detection -- A Comprehensive Review

Yuki Hagiwara, Octavia-Andreaa Ciora, Maureen Monnet, Gino Lancho, Jeanette Miriam Lorenz

The diagnosis of glaucoma plays a critical role in the management and treatment of this vision-threatening disease. Glaucoma is a group of eye diseases that cause blindness by damaging the optic nerve at the back of the eye. Often called "silent thief of sight", it exhibits no symptoms during the early stages. Therefore, early detection is crucial to prevent vision loss. With the rise of Artificial Intelligence (AI), particularly Deep Learning (DL) techniques, Computer-Aided Diagnosis (CADx) systems have emerged as promising tools to assist clinicians in accurately diagnosing glaucoma early. This paper aims to provide a comprehensive overview of AI techniques utilized in CADx systems for glaucoma diagnosis. Through a detailed analysis of current literature, we identify key gaps and challenges in these systems, emphasizing the need for improved safety, reliability, interpretability, and explainability. By identifying research gaps, we aim to advance the field of CADx systems especially for the early diagnosis of glaucoma, in order to prevent any potential loss of vision.

☆ MBPU: A Plug-and-Play State Space Model for Point Cloud Upsamping with Fast Point Rendering

Jiayi Song, Weidong Yang, Zhijun Li, Wen-Ming Chen, Ben Fei

The task of point cloud upsampling (PCU) is to generate dense and uniform point clouds from sparse input captured by 3D sensors like LiDAR, holding potential applications in real yet is still a challenging task. Existing deep learning-based methods have shown significant achievements in this field. However, they still face limitations in effectively handling long sequences and addressing the issue of shrinkage artifacts around the surface of the point cloud. Inspired by the newly proposed Mamba, in this paper, we introduce a network named MBPU built on top of the Mamba architecture, which performs well in long sequence modeling, especially for large-scale point cloud upsampling, and achieves fast convergence speed. Moreover, MBPU is an arbitrary-scale upsampling framework as the predictor of point distance in the point refinement phase. At the same time, we simultaneously predict the 3D position shift and 1D point-to-point distance as regression quantities to constrain the global features while ensuring the accuracy of local details. We also introduce a fast differentiable renderer to further enhance the fidelity of the upsampled point cloud and reduce artifacts. It is noted that, by the merits of our fast point rendering, MBPU yields high-quality upsampled point clouds by effectively eliminating surface noise. Extensive experiments have demonstrated that our MBPU outperforms other off-the-shelf methods in terms of point cloud upsampling, especially for large-scale point clouds.

☆ Focus on BEV: Self-calibrated Cycle View Transformation for Monocular Birds-Eye-View Segmentation

Jiawei Zhao, Qixing Jiang, Xuede Li, Junfeng Luo

Birds-Eye-View (BEV) segmentation aims to establish a spatial mapping from the perspective view to the top view and estimate the semantic maps from monocular images. Recent studies have encountered difficulties in view transformation due to the disruption of BEV-agnostic features in image space. To tackle this issue, we propose a novel FocusBEV framework consisting of $(i)$ a self-calibrated cross view transformation module to suppress the BEV-agnostic image areas and focus on the BEV-relevant areas in the view transformation stage, $(ii)$ a plug-and-play ego-motion-based temporal fusion module to exploit the spatiotemporal structure consistency in BEV space with a memory bank, and $(iii)$ an occupancy-agnostic IoU loss to mitigate both semantic and positional uncertainties. Experimental evidence demonstrates that our approach achieves new state-of-the-art on two popular benchmarks,\ie, 29.2\% mIoU on nuScenes and 35.2\% mIoU on Argoverse.

☆ GReFEL: Geometry-Aware Reliable Facial Expression Learning under Bias and Imbalanced Data Distribution ACCV 2024

Azmine Toushik Wasi, Taki Hasan Rafi, Raima Islam, Karlo Serbetar, Dong Kyu Chae

Reliable facial expression learning (FEL) involves the effective learning of distinctive facial expression characteristics for more reliable, unbiased and accurate predictions in real-life settings. However, current systems struggle with FEL tasks because of the variance in people's facial expressions due to their unique facial structures, movements, tones, and demographics. Biased and imbalanced datasets compound this challenge, leading to wrong and biased prediction labels. To tackle these, we introduce GReFEL, leveraging Vision Transformers and a facial geometry-aware anchor-based reliability balancing module to combat imbalanced data distributions, bias, and uncertainty in facial expression learning. Integrating local and global data with anchors that learn different facial data points and structural features, our approach adjusts biased and mislabeled emotions caused by intra-class disparity, inter-class similarity, and scale sensitivity, resulting in comprehensive, accurate, and reliable facial expression predictions. Our model outperforms current state-of-the-art methodologies, as demonstrated by extensive experiments on various datasets.

comment: ACCV 2024. Extended version of ARBEx (arXiv:2305.01486)

☆ Mitigating Object Hallucination via Concentric Causal Attention NeurIPS 2024

Yun Xing, Yiheng Li, Ivan Laptev, Shijian Lu

Recent Large Vision Language Models (LVLMs) present remarkable zero-shot conversational and reasoning capabilities given multimodal queries. Nevertheless, they suffer from object hallucination, a phenomenon where LVLMs are prone to generate textual responses not factually aligned with image inputs. Our pilot study reveals that object hallucination is closely tied with Rotary Position Encoding (RoPE), a widely adopted positional dependency modeling design in existing LVLMs. Due to the long-term decay in RoPE, LVLMs tend to hallucinate more when relevant visual cues are distant from instruction tokens in the multimodal input sequence. Additionally, we observe a similar effect when reversing the sequential order of visual tokens during multimodal alignment. Our tests indicate that long-term decay in RoPE poses challenges to LVLMs while capturing visual-instruction interactions across long distances. We propose Concentric Causal Attention (CCA), a simple yet effective positional alignment strategy that mitigates the impact of RoPE long-term decay in LVLMs by naturally reducing relative distance between visual and instruction tokens. With CCA, visual tokens can better interact with instruction tokens, thereby enhancing model's perception capability and alleviating object hallucination. Without bells and whistles, our positional alignment method surpasses existing hallucination mitigation strategies by large margins on multiple object hallucination benchmarks.

comment: To appear at NeurIPS 2024. Code is available at https://github.com/xing0047/cca-llava

☆ Are Large-scale Soft Labels Necessary for Large-scale Dataset Distillation?

Lingao Xiao, Yang He

In ImageNet-condensation, the storage for auxiliary soft labels exceeds that of the condensed dataset by over 30 times. However, are large-scale soft labels necessary for large-scale dataset distillation? In this paper, we first discover that the high within-class similarity in condensed datasets necessitates the use of large-scale soft labels. This high within-class similarity can be attributed to the fact that previous methods use samples from different classes to construct a single batch for batch normalization (BN) matching. To reduce the within-class similarity, we introduce class-wise supervision during the image synthesizing process by batching the samples within classes, instead of across classes. As a result, we can increase within-class diversity and reduce the size of required soft labels. A key benefit of improved image diversity is that soft label compression can be achieved through simple random pruning, eliminating the need for complex rule-based strategies. Experiments validate our discoveries. For example, when condensing ImageNet-1K to 200 images per class, our approach compresses the required soft labels from 113 GB to 2.8 GB (40x compression) with a 2.6% performance gain. Code is available at: https://github.com/he-y/soft-label-pruning-for-dataset-distillation

comment: Accepted by Neurips 2024

☆ Leveraging CORAL-Correlation Consistency Network for Semi-Supervised Left Atrium MRI Segmentation

Xinze Li, Runlin Huang, Zhenghao Wu, Bohan Yang, Wentao Fan, Chengzhang Zhu, Weifeng Su

Semi-supervised learning (SSL) has been widely used to learn from both a few labeled images and many unlabeled images to overcome the scarcity of labeled samples in medical image segmentation. Most current SSL-based segmentation methods use pixel values directly to identify similar features in labeled and unlabeled data. They usually fail to accurately capture the intricate attachment structures in the left atrium, such as the areas of inconsistent density or exhibit outward curvatures, adding to the complexity of the task. In this paper, we delve into this issue and introduce an effective solution, CORAL(Correlation-Aligned)-Correlation Consistency Network (CORN), to capture the global structure shape and local details of Left Atrium. Diverging from previous methods focused on each local pixel value, the CORAL-Correlation Consistency Module (CCM) in the CORN leverages second-order statistical information to capture global structural features by minimizing the distribution discrepancy between labeled and unlabeled samples in feature space. Yet, direct construction of features from unlabeled data frequently results in ``Sample Selection Bias'', leading to flawed supervision. We thus further propose the Dynamic Feature Pool (DFP) for the CCM, which utilizes a confidence-based filtering strategy to remove incorrectly selected features and regularize both teacher and student models by constraining the similarity matrix to be consistent. Extensive experiments on the Left Atrium dataset have shown that the proposed CORN outperforms previous state-of-the-art semi-supervised learning methods.

comment: 5 pages, 3 figures, Accepted by 2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM 2024)

☆ Hybrid Architecture for Real-Time Video Anomaly Detection: Integrating Spatial and Temporal Analysis

Fabien Poirier

We propose a new architecture for real-time anomaly detection in video data, inspired by human behavior by combining spatial and temporal analyses. This approach uses two distinct models: for temporal analysis, a recurrent convolutional network (CNN + RNN) is employed, associating VGG19 and a GRU to process video sequences. Regarding spatial analysis, it is performed using YOLOv7 to analyze individual images. These two analyses can be carried out either in parallel, with a final prediction that combines the results of both analyses, or in series, where the spatial analysis enriches the data before the temporal analysis. In this article, we will compare these two architectural configurations with each other, to evaluate the effectiveness of our hybrid approach in video anomaly detection.

☆ Seismic Phase Picking

Yuchen Wang, Ruihuan Wang

Seismic phase picking, which aims to determine the arrival time of P- and S-waves according to seismic waveforms, is fundamental to earthquake monitoring. Generally, manual phase picking is trustworthy, but with the increasing number of worldwide stations and seismic monitors, it becomes more challenging for human to complete the task comprehensively. In this work, we explore multiple ways to do automatic phase picking, including traditional and learning-based methods.

☆ TexPro: Text-guided PBR Texturing with Procedural Material Modeling

Ziqiang Dang, Wenqi Dong, Zesong Yang, Bangbang Yang, Liang Li, Yuewen Ma, Zhaopeng Cui

In this paper, we present TexPro, a novel method for high-fidelity material generation for input 3D meshes given text prompts. Unlike existing text-conditioned texture generation methods that typically generate RGB textures with baked lighting, TexPro is able to produce diverse texture maps via procedural material modeling, which enables physical-based rendering, relighting, and additional benefits inherent to procedural materials. Specifically, we first generate multi-view reference images given the input textual prompt by employing the latest text-to-image model. We then derive texture maps through a rendering-based optimization with recent differentiable procedural materials. To this end, we design several techniques to handle the misalignment between the generated multi-view images and 3D meshes, and introduce a novel material agent that enhances material classification and matching by exploring both part-level understanding and object-aware material reasoning. Experiments demonstrate the superiority of the proposed method over existing SOTAs and its capability of relighting.

comment: In submission. Supplementary material is included at the end of the main paper (5 pages, 2 figures)

☆ Foundation Models for Slide-level Cancer Subtyping in Digital Pathology SC

Pablo Meseguer, Rocío del Amor, Adrian Colomer, Valery Naranjo

Since the emergence of the ImageNet dataset, the pretraining and fine-tuning approach has become widely adopted in computer vision due to the ability of ImageNet-pretrained models to learn a wide variety of visual features. However, a significant challenge arises when adapting these models to domain-specific fields, such as digital pathology, due to substantial gaps between domains. To address this limitation, foundation models (FM) have been trained on large-scale in-domain datasets to learn the intricate features of histopathology images. In cancer diagnosis, whole-slide image (WSI) prediction is essential for patient prognosis, and multiple instance learning (MIL) has been implemented to handle the giga-pixel size of WSI. As MIL frameworks rely on patch-level feature aggregation, this work aims to compare the performance of various feature extractors developed under different pretraining strategies for cancer subtyping on WSI under a MIL framework. Results demonstrate the ability of foundation models to surpass ImageNet-pretrained models for the prediction of six skin cancer subtypes

comment: Manuscript accepted for oral presentation at Decision Science Allieance -INternational Summer Conference (DSA-ISC) 2024 held on Valencia, Spain

☆ Distributed Learning for UAV Swarms

Chen Hu, Hanchi Ren, Jingjing Deng, Xianghua Xie

Unmanned Aerial Vehicle (UAV) swarms are increasingly deployed in dynamic, data-rich environments for applications such as environmental monitoring and surveillance. These scenarios demand efficient data processing while maintaining privacy and security, making Federated Learning (FL) a promising solution. FL allows UAVs to collaboratively train global models without sharing raw data, but challenges arise due to the non-Independent and Identically Distributed (non-IID) nature of the data collected by UAVs. In this study, we show an integration of the state-of-the-art FL methods to UAV Swarm application and invetigate the performance of multiple aggregation methods (namely FedAvg, FedProx, FedOpt, and MOON) with a particular focus on tackling non-IID on a variety of datasets, specifically MNIST for baseline performance, CIFAR10 for natural object classification, EuroSAT for environment monitoring, and CelebA for surveillance. These algorithms were selected to cover improved techniques on both client-side updates and global aggregation. Results show that while all algorithms perform comparably on IID data, their performance deteriorates significantly under non-IID conditions. FedProx demonstrated the most stable overall performance, emphasising the importance of regularising local updates in non-IID environments to mitigate drastic deviations in local models.

☆ MI-VisionShot: Few-shot adaptation of vision-language models for slide-level classification of histopathological images

Pablo Meseguer, Rocío del Amor, Valery Naranjo

Vision-language supervision has made remarkable strides in learning visual representations from textual guidance. In digital pathology, vision-language models (VLM), pre-trained on curated datasets of histological image-captions, have been adapted to downstream tasks, such as region of interest classification. Zero-shot transfer for slide-level prediction has been formulated by MI-Zero, but it exhibits high variability depending on the textual prompts. Inspired by prototypical learning, we propose MI-VisionShot, a training-free adaptation method on top of VLMs to predict slide-level labels in few-shot learning scenarios. Our framework takes advantage of the excellent representation learning of VLM to create prototype-based classifiers under a multiple-instance setting by retrieving the most discriminative patches within each slide. Experimentation through different settings shows the ability of MI-VisionShot to surpass zero-shot transfer with lower variability, even in low-shot scenarios. Code coming soon at thttps://github.com/cvblab/MIVisionShot.

comment: Manuscript accepted for oral presentation at KES-InnovationInMedicine 2024 held on Madeira, Portugal

☆ Visual Motif Identification: Elaboration of a Curated Comparative Dataset and Classification Methods ECCV 2024

Adam Phillips, Daniel Grandes Rodriguez, Miriam Sánchez-Manzano, Alan Salvadó, Manuel Garin, Gloria Haro, Coloma Ballester

In cinema, visual motifs are recurrent iconographic compositions that carry artistic or aesthetic significance. Their use throughout the history of visual arts and media is interesting to researchers and filmmakers alike. Our goal in this work is to recognise and classify these motifs by proposing a new machine learning model that uses a custom dataset to that end. We show how features extracted from a CLIP model can be leveraged by using a shallow network and an appropriate loss to classify images into 20 different motifs, with surprisingly good results: an $F_1$-score of 0.91 on our test set. We also present several ablation studies justifying the input features, architecture and hyperparameters used.

comment: 17 pages, 11 figures, one table, to be published in the conference proceedings of ECCV 2024

☆ R2I-rPPG: A Robust Region of Interest Selection Method for Remote Photoplethysmography to Extract Heart Rate

Sandeep Nagar, Mark Hasegawa-Johnson, David G. Beiser, Narendra Ahuja

The COVID-19 pandemic has underscored the need for low-cost, scalable approaches to measuring contactless vital signs, either during initial triage at a healthcare facility or virtual telemedicine visits. Remote photoplethysmography (rPPG) can accurately estimate heart rate (HR) when applied to close-up videos of healthy volunteers in well-lit laboratory settings. However, results from such highly optimized laboratory studies may not be readily translated to healthcare settings. One significant barrier to the practical application of rPPG in health care is the accurate localization of the region of interest (ROI). Clinical or telemedicine visits may involve sub-optimal lighting, movement artifacts, variable camera angle, and subject distance. This paper presents an rPPG ROI selection method based on 3D facial landmarks and patient head yaw angle. We then demonstrate the robustness of this ROI selection method when coupled to the Plane-Orthogonal-to-Skin (POS) rPPG method when applied to videos of patients presenting to an Emergency Department for respiratory complaints. Our results demonstrate the effectiveness of our proposed approach in improving the accuracy and robustness of rPPG in a challenging clinical environment.

comment: preprint

☆ Random Token Fusion for Multi-View Medical Diagnosis NeurIPS 2024

Jingyu Guo, Christos Matsoukas, Fredrik Strand, Kevin Smith

In multi-view medical diagnosis, deep learning-based models often fuse information from different imaging perspectives to improve diagnostic performance. However, existing approaches are prone to overfitting and rely heavily on view-specific features, which can lead to trivial solutions. In this work, we introduce Random Token Fusion (RTF), a novel technique designed to enhance multi-view medical image analysis using vision transformers. By integrating randomness into the feature fusion process during training, RTF addresses the issue of overfitting and enhances the robustness and accuracy of diagnostic models without incurring any additional cost at inference. We validate our approach on standard mammography and chest X-ray benchmark datasets. Through extensive experiments, we demonstrate that RTF consistently improves the performance of existing fusion methods, paving the way for a new generation of multi-view medical foundation models.

comment: Originally published at the NeurIPS 2024 Workshop on Advancements In Medical Foundation Models: Explainability, Robustness, Security, and Beyond (AIM-FM)

☆ LiOn-XA: Unsupervised Domain Adaptation via LiDAR-Only Cross-Modal Adversarial Training IROS2024

Thomas Kreutz, Jens Lemke, Max Mühlhäuser, Alejandro Sanchez Guinea

In this paper, we propose LiOn-XA, an unsupervised domain adaptation (UDA) approach that combines LiDAR-Only Cross-Modal (X) learning with Adversarial training for 3D LiDAR point cloud semantic segmentation to bridge the domain gap arising from environmental and sensor setup changes. Unlike existing works that exploit multiple data modalities like point clouds and RGB image data, we address UDA in scenarios where RGB images might not be available and show that two distinct LiDAR data representations can learn from each other for UDA. More specifically, we leverage 3D voxelized point clouds to preserve important geometric structure in combination with 2D projection-based range images that provide information such as object orientations or surfaces. To further align the feature space between both domains, we apply adversarial training using both features and predictions of both 2D and 3D neural networks. Our experiments on 3 real-to-real adaptation scenarios demonstrate the effectiveness of our approach, achieving new state-of-the-art performance when compared to previous uni- and multi-model UDA methods. Our source code is publicly available at https://github.com/JensLe97/lion-xa.

comment: Preprint, Paper has been accepted at IROS2024

☆ LiMTR: Time Series Motion Prediction for Diverse Road Users through Multimodal Feature Integration NeurIPS 2024

Camiel Oerlemans, Bram Grooten, Michiel Braat, Alaa Alassi, Emilia Silvas, Decebal Constantin Mocanu

Predicting the behavior of road users accurately is crucial to enable the safe operation of autonomous vehicles in urban or densely populated areas. Therefore, there has been a growing interest in time series motion prediction research, leading to significant advancements in state-of-the-art techniques in recent years. However, the potential of using LiDAR data to capture more detailed local features, such as a person's gaze or posture, remains largely unexplored. To address this, we develop a novel multimodal approach for motion prediction based on the PointNet foundation model architecture, incorporating local LiDAR features. Evaluation on the Waymo Open Dataset shows a performance improvement of 6.20% and 1.58% in minADE and mAP respectively, when integrated and compared with the previous state-of-the-art MTR. We open-source the code of our LiMTR model.

comment: Accepted at the NeurIPS 2024 workshop Time Series in the Age of Large Models. Code available at https://github.com/Cing2/LiMTR

☆ Kaninfradet3D:A Road-side Camera-LiDAR Fusion 3D Perception Model based on Nonlinear Feature Extraction and Intrinsic Correlation

Pei Liu, Nanfang Zheng, Yiqun Li, Junlan Chen, Ziyuan Pu

With the development of AI-assisted driving, numerous methods have emerged for ego-vehicle 3D perception tasks, but there has been limited research on roadside perception. With its ability to provide a global view and a broader sensing range, the roadside perspective is worth developing. LiDAR provides precise three-dimensional spatial information, while cameras offer semantic information. These two modalities are complementary in 3D detection. However, adding camera data does not increase accuracy in some studies since the information extraction and fusion procedure is not sufficiently reliable. Recently, Kolmogorov-Arnold Networks (KANs) have been proposed as replacements for MLPs, which are better suited for high-dimensional, complex data. Both the camera and the LiDAR provide high-dimensional information, and employing KANs should enhance the extraction of valuable features to produce better fusion outcomes. This paper proposes Kaninfradet3D, which optimizes the feature extraction and fusion modules. To extract features from complex high-dimensional data, the model's encoder and fuser modules were improved using KAN Layers. Cross-attention was applied to enhance feature fusion, and visual comparisons verified that camera features were more evenly integrated. This addressed the issue of camera features being abnormally concentrated, negatively impacting fusion. Compared to the benchmark, our approach shows improvements of +9.87 mAP and +10.64 mAP in the two viewpoints of the TUMTraf Intersection Dataset and an improvement of +1.40 mAP in the roadside end of the TUMTraf V2X Cooperative Perception Dataset. The results indicate that Kaninfradet3D can effectively fuse features, demonstrating the potential of applying KANs in roadside perception tasks.

☆ FusionLungNet: Multi-scale Fusion Convolution with Refinement Network for Lung CT Image Segmentation

Sadjad Rezvani, Mansoor Fateh, Yeganeh Jalali, Amirreza Fateh

Early detection of lung cancer is crucial as it increases the chances of successful treatment. Automatic lung image segmentation assists doctors in identifying diseases such as lung cancer, COVID-19, and respiratory disorders. However, lung segmentation is challenging due to overlapping features like vascular and bronchial structures, along with pixel-level fusion of brightness, color, and texture. New lung segmentation methods face difficulties in identifying long-range relationships between image components, reliance on convolution operations that may not capture all critical features, and the complex structures of the lungs. Furthermore, semantic gaps between feature maps can hinder the integration of relevant information, reducing model accuracy. Skip connections can also limit the decoder's access to complete information, resulting in partial information loss during encoding. To overcome these challenges, we propose a hybrid approach using the FusionLungNet network, which has a multi-level structure with key components, including the ResNet-50 encoder, Channel-wise Aggregation Attention (CAA) module, Multi-scale Feature Fusion (MFF) block, self refinement (SR) module, and multiple decoders. The refinement sub-network uses convolutional neural networks for image post-processing to improve quality. Our method employs a combination of loss functions, including SSIM, IOU, and focal loss, to optimize image reconstruction quality. We created and publicly released a new dataset for lung segmentation called LungSegDB, including 1800 CT images from the LIDC-IDRI dataset (dataset version 1) and 700 images from the Chest CT Cancer Images from Kaggle dataset (dataset version 2). Our method achieved an IOU score of 98.04, outperforming existing methods and demonstrating significant improvements in segmentation accuracy. https://github.com/sadjadrz/FusionLungNet

☆ Data-Efficient CLIP-Powered Dual-Branch Networks for Source-Free Unsupervised Domain Adaptation

Yongguang Li, Yueqi Cao, Jindong Li, Qi Wang, Shengsheng Wang

Source-Free Unsupervised Domain Adaptation (SF-UDA) aims to transfer a model's performance from a labeled source domain to an unlabeled target domain without direct access to source samples, addressing data privacy issues. However, most existing SF-UDA approaches assume the availability of abundant source domain samples, which is often impractical due to the high cost of data annotation. In this paper, we explore a more challenging scenario where direct access to source domain samples is restricted, and the source domain contains only a few samples. To tackle the dual challenges of limited source data and privacy concerns, we introduce a data-efficient, CLIP-powered dual-branch network (CDBN in short). We design a cross-modal dual-branch network that integrates source domain class semantics into the unsupervised fine-tuning of the target domain. It preserves the class information from the source domain while enhancing the model's generalization to the target domain. Additionally, we propose an unsupervised optimization strategy driven by accurate classification and diversity, which aims to retain the classification capability learned from the source domain while producing more confident and diverse predictions in the target domain. Extensive experiments across 31 transfer tasks on 7 public datasets demonstrate that our approach achieves state-of-the-art performance compared to existing methods.

☆ Assisted Physical Interaction: Autonomous Aerial Robots with Neural Network Detection, Navigation, and Safety Layers

Andrea Berra, Viswa Narayanan Sankaranarayanan, Achilleas Santi Seisa, Julien Mellet, Udayanga G. W. K. N. Gamage, Sumeet Gajanan Satpute, Fabio Ruggiero, Vincenzo Lippiello, Silvia Tolu, Matteo Fumagalli, George Nikolakopoulos, Miguel Ángel Trujillo Soto, Guillermo Heredia

The paper introduces a novel framework for safe and autonomous aerial physical interaction in industrial settings. It comprises two main components: a neural network-based target detection system enhanced with edge computing for reduced onboard computational load, and a control barrier function (CBF)-based controller for safe and precise maneuvering. The target detection system is trained on a dataset under challenging visual conditions and evaluated for accuracy across various unseen data with changing lighting conditions. Depth features are utilized for target pose estimation, with the entire detection framework offloaded into low-latency edge computing. The CBF-based controller enables the UAV to converge safely to the target for precise contact. Simulated evaluations of both the controller and target detection are presented, alongside an analysis of real-world detection performance.

comment: 8 pages,14 figures, ICUAS 2024

☆ Habaek: High-performance water segmentation through dataset expansion and inductive bias optimization

Hanseon Joo, Eunji Lee, Minjong Cheon

Water segmentation is critical to disaster response and water resource management. Authorities may employ high-resolution photography to monitor rivers, lakes, and reservoirs, allowing for more proactive management in agriculture, industry, and conservation. Deep learning has improved flood monitoring by allowing models like CNNs, U-Nets, and transformers to handle large volumes of satellite and aerial data. However, these models usually have significant processing requirements, limiting their usage in real-time applications. This research proposes upgrading the SegFormer model for water segmentation by data augmentation with datasets such as ADE20K and RIWA to boost generalization. We examine how inductive bias affects attention-based models and discover that SegFormer performs better on bigger datasets. To further demonstrate the function of data augmentation, Low-Rank Adaptation (LoRA) is used to lower processing complexity while preserving accuracy. We show that the suggested Habaek model outperforms current models in segmentation, with an Intersection over Union (IoU) ranging from 0.91986 to 0.94397. In terms of F1-score, recall, accuracy, and precision, Habaek performs better than rival models, indicating its potential for real-world applications. This study highlights the need to enhance structures and include datasets for effective water segmentation.

☆ WildOcc: A Benchmark for Off-Road 3D Semantic Occupancy Prediction

Heng Zhai, Jilin Mei, Chen Min, Liang Chen, Fangzhou Zhao, Yu Hu

3D semantic occupancy prediction is an essential part of autonomous driving, focusing on capturing the geometric details of scenes. Off-road environments are rich in geometric information, therefore it is suitable for 3D semantic occupancy prediction tasks to reconstruct such scenes. However, most of researches concentrate on on-road environments, and few methods are designed for off-road 3D semantic occupancy prediction due to the lack of relevant datasets and benchmarks. In response to this gap, we introduce WildOcc, to our knowledge, the first benchmark to provide dense occupancy annotations for off-road 3D semantic occupancy prediction tasks. A ground truth generation pipeline is proposed in this paper, which employs a coarse-to-fine reconstruction to achieve a more realistic result. Moreover, we introduce a multi-modal 3D semantic occupancy prediction framework, which fuses spatio-temporal information from multi-frame images and point clouds at voxel level. In addition, a cross-modality distillation function is introduced, which transfers geometric knowledge from point clouds to image features.

☆ An Efficient System for Automatic Map Storytelling -- A Case Study on Historical Maps

Ziyi Liu, Claudio Affolter, Sidi Wu, Yizi Chen, Lorenz Hurni

Historical maps provide valuable information and knowledge about the past. However, as they often feature non-standard projections, hand-drawn styles, and artistic elements, it is challenging for non-experts to identify and interpret them. While existing image captioning methods have achieved remarkable success on natural images, their performance on maps is suboptimal as maps are underrepresented in their pre-training process. Despite the recent advance of GPT-4 in text recognition and map captioning, it still has a limited understanding of maps, as its performance wanes when texts (e.g., titles and legends) in maps are missing or inaccurate. Besides, it is inefficient or even impractical to fine-tune the model with users' own datasets. To address these problems, we propose a novel and lightweight map-captioning counterpart. Specifically, we fine-tune the state-of-the-art vision-language model CLIP to generate captions relevant to historical maps and enrich the captions with GPT-3.5 to tell a brief story regarding where, what, when and why of a given map. We propose a novel decision tree architecture to only generate captions relevant to the specified map type. Our system shows invariance to text alterations in maps. The system can be easily adapted and extended to other map types and scaled to a larger map captioning system. The code is open-sourced at https://github.com/claudaff/automatic-map-storytelling.

☆ Reducing Hallucinations in Vision-Language Models via Latent Space Steering

Sheng Liu, Haotian Ye, James Zou

Hallucination poses a challenge to the deployment of large vision-language models (LVLMs) in applications. Unlike in large language models (LLMs), hallucination in LVLMs often arises from misalignments between visual inputs and textual outputs. This paper investigates the underlying mechanisms of hallucination, focusing on the unique structure of LVLMs that distinguishes them from large language models (LLMs). We identify that hallucinations often arise from the sensitivity of text decoders to vision inputs, a natural phenomenon when image encoders and text decoders are pre-trained separately. Inspired by this, we introduce Visual and Textual Intervention (VTI), a novel technique designed to reduce hallucinations by steering latent space representations during inference to enhance the stability of vision features. As a task-agnostic test-time intervention, VTI can be easily applied to any problem without additional cost. Extensive experiments demonstrate that it can effectively reduce hallucinations and outperform baseline methods across multiple metrics, highlighting the critical role of vision feature stability in LVLMs.

comment: 21 pages

★ Generalizing Motion Planners with Mixture of Experts for Autonomous Driving

Qiao Sun, Huimin Wang, Jiahao Zhan, Fan Nie, Xin Wen, Leimeng Xu, Kun Zhan, Peng Jia, Xianpeng Lang, Hang Zhao

Large real-world driving datasets have sparked significant research into various aspects of data-driven motion planners for autonomous driving. These include data augmentation, model architecture, reward design, training strategies, and planner pipelines. These planners promise better generalizations on complicated and few-shot cases than previous methods. However, experiment results show that many of these approaches produce limited generalization abilities in planning performance due to overly complex designs or training paradigms. In this paper, we review and benchmark previous methods focusing on generalizations. The experimental results indicate that as models are appropriately scaled, many design elements become redundant. We introduce StateTransformer-2 (STR2), a scalable, decoder-only motion planner that uses a Vision Transformer (ViT) encoder and a mixture-of-experts (MoE) causal Transformer architecture. The MoE backbone addresses modality collapse and reward balancing by expert routing during training. Extensive experiments on the NuPlan dataset show that our method generalizes better than previous approaches across different test sets and closed-loop simulations. Furthermore, we assess its scalability on billions of real-world urban driving scenarios, demonstrating consistent accuracy improvements as both data and model size grow.

comment: 7 pages, 3 figures

☆ Learning to Synthesize Graphics Programs for Geometric Artworks ICPR 2024

Qi Bing, Chaoyi Zhang, Weidong Cai

Creating and understanding art has long been a hallmark of human ability. When presented with finished digital artwork, professional graphic artists can intuitively deconstruct and replicate it using various drawing tools, such as the line tool, paint bucket, and layer features, including opacity and blending modes. While most recent research in this field has focused on art generation, proposing a range of methods, these often rely on the concept of artwork being represented as a final image. To bridge the gap between pixel-level results and the actual drawing process, we present an approach that treats a set of drawing tools as executable programs. This method predicts a sequence of steps to achieve the final image, allowing for understandable and resolution-independent reproductions under the usage of a set of drawing commands. Our experiments demonstrate that our program synthesizer, Art2Prog, can comprehensively understand complex input images and reproduce them using high-quality executable programs. The experimental results evidence the potential of machines to grasp higher-level information from images and generate compact program-level descriptions.

comment: ICPR 2024

☆ Improving Instance Optimization in Deformable Image Registration with Gradient Projection

Yi Zhang, Yidong Zhao, Qian Tao

Deformable image registration is inherently a multi-objective optimization (MOO) problem, requiring a delicate balance between image similarity and deformation regularity. These conflicting objectives often lead to poor optimization outcomes, such as being trapped in unsatisfactory local minima or experiencing slow convergence. Deep learning methods have recently gained popularity in this domain due to their efficiency in processing large datasets and achieving high accuracy. However, they often underperform during test time compared to traditional optimization techniques, which further explore iterative, instance-specific gradient-based optimization. This performance gap is more pronounced when a distribution shift between training and test data exists. To address this issue, we focus on the instance optimization (IO) paradigm, which involves additional optimization for test-time instances based on a pre-trained model. IO effectively combines the generalization capabilities of deep learning with the fine-tuning advantages of instance-specific optimization. Within this framework, we emphasize the use of gradient projection to mitigate conflicting updates in MOO. This technique projects conflicting gradients into a common space, better aligning the dual objectives and enhancing optimization stability. We validate our method using a state-of-the-art foundation model on the 3D Brain inter-subject registration task (LUMIR) from the Learn2Reg 2024 Challenge. Our results show significant improvements over standard gradient descent, leading to more accurate and reliable registration results.

comment: L2R 2024 Challenge Paper

☆ How Important are Data Augmentations to Close the Domain Gap for Object Detection in Orbit?

Maximilian Ulmer, Leonard Klüpfel, Maximilian Durner, Rudolph Triebel

We investigate the efficacy of data augmentations to close the domain gap in spaceborne computer vision, crucial for autonomous operations like on-orbit servicing. As the use of computer vision in space increases, challenges such as hostile illumination and low signal-to-noise ratios significantly hinder performance. While learning-based algorithms show promising results, their adoption is limited by the need for extensive annotated training data and the domain gap that arises from differences between synthesized and real-world imagery. This study explores domain generalization in terms of data augmentations -- classical color and geometric transformations, corruptions, and noise -- to enhance model performance across the domain gap. To this end, we conduct an large scale experiment using a hyperparameter optimization pipeline that samples hundreds of different configurations and searches for the best set to bridge the domain gap. As a reference task, we use 2D object detection and evaluate on the SPEED+ dataset that contains real hardware-in-the-loop satellite images in its test set. Moreover, we evaluate four popular object detectors, including Mask R-CNN, Faster R-CNN, YOLO-v7, and the open set detector GroundingDINO, and highlight their trade-offs between performance, inference speed, and training time. Our results underscore the vital role of data augmentations in bridging the domain gap, improving model performance, robustness, and reliability for critical space applications. As a result, we propose two novel data augmentations specifically developed to emulate the visual effects observed in orbital imagery. We conclude by recommending the most effective augmentations for advancing computer vision in challenging orbital environments. Code for training detectors and hyperparameter search will be made publicly available.

☆ DeepIcon: A Hierarchical Network for Layer-wise Icon Vectorization

Qi Bing, Chaoyi Zhang, Weidong Cai

In contrast to the well-established technique of rasterization, vectorization of images poses a significant challenge in the field of computer graphics. Recent learning-based methods for converting raster images to vector formats frequently suffer from incomplete shapes, redundant path prediction, and a lack of accuracy in preserving the semantics of the original content. These shortcomings severely hinder the utility of these methods for further editing and manipulation of images. To address these challenges, we present DeepIcon, a novel hierarchical image vectorization network specifically tailored for generating variable-length icon vector graphics based on the raster image input. Our experimental results indicate that DeepIcon can efficiently produce Scalable Vector Graphics (SVGs) directly from raster images, bypassing the need for a differentiable rasterizer while also demonstrating a profound understanding of the image contents.

comment: Accepted as Oral Presentation at DICTA 2024

☆ Unleashing the Potential of Vision-Language Pre-Training for 3D Zero-Shot Lesion Segmentation via Mask-Attribute Alignment

Yankai Jiang, Wenhui Lei, Xiaofan Zhang, Shaoting Zhang

Recent advancements in medical vision-language pre-training models have driven significant progress in zero-shot disease recognition. However, transferring image-level knowledge to pixel-level tasks, such as lesion segmentation in 3D CT scans, remains a critical challenge. Due to the complexity and variability of pathological visual characteristics, existing methods struggle to align fine-grained lesion features not encountered during training with disease-related textual representations. In this paper, we present Malenia, a novel multi-scale lesion-level mask-attribute alignment framework, specifically designed for 3D zero-shot lesion segmentation. Malenia improves the compatibility between mask representations and their associated elemental attributes, explicitly linking the visual features of unseen lesions with the extensible knowledge learned from previously seen ones. Furthermore, we design a Cross-Modal Knowledge Injection module to enhance both visual and textual features with mutually beneficial information, effectively guiding the generation of segmentation results. Comprehensive experiments across three datasets and 12 lesion categories validate the superior performance of Malenia. Codes will be publicly available.

☆ ViMoE: An Empirical Study of Designing Vision Mixture-of-Experts

Xumeng Han, Longhui Wei, Zhiyang Dou, Zipeng Wang, Chenhui Qiang, Xin He, Yingfei Sun, Zhenjun Han, Qi Tian

Mixture-of-Experts (MoE) models embody the divide-and-conquer concept and are a promising approach for increasing model capacity, demonstrating excellent scalability across multiple domains. In this paper, we integrate the MoE structure into the classic Vision Transformer (ViT), naming it ViMoE, and explore the potential of applying MoE to vision through a comprehensive study on image classification. However, we observe that the performance is sensitive to the configuration of MoE layers, making it challenging to obtain optimal results without careful design. The underlying cause is that inappropriate MoE layers lead to unreliable routing and hinder experts from effectively acquiring helpful knowledge. To address this, we introduce a shared expert to learn and capture common information, serving as an effective way to construct stable ViMoE. Furthermore, we demonstrate how to analyze expert routing behavior, revealing which MoE layers are capable of specializing in handling specific information and which are not. This provides guidance for retaining the critical layers while removing redundancies, thereby advancing ViMoE to be more efficient without sacrificing accuracy. We aspire for this work to offer new insights into the design of vision MoE models and provide valuable empirical guidance for future research.

★ Object-Centric Temporal Consistency via Conditional Autoregressive Inductive Biases

Cristian Meo, Akihiro Nakano, Mircea Lică, Aniket Didolkar, Masahiro Suzuki, Anirudh Goyal, Mengmi Zhang, Justin Dauwels, Yutaka Matsuo, Yoshua Bengio

Unsupervised object-centric learning from videos is a promising approach towards learning compositional representations that can be applied to various downstream tasks, such as prediction and reasoning. Recently, it was shown that pretrained Vision Transformers (ViTs) can be useful to learn object-centric representations on real-world video datasets. However, while these approaches succeed at extracting objects from the scenes, the slot-based representations fail to maintain temporal consistency across consecutive frames in a video, i.e. the mapping of objects to slots changes across the video. To address this, we introduce Conditional Autoregressive Slot Attention (CA-SA), a framework that enhances the temporal consistency of extracted object-centric representations in video-centric vision tasks. Leveraging an autoregressive prior network to condition representations on previous timesteps and a novel consistency loss function, CA-SA predicts future slot representations and imposes consistency across frames. We present qualitative and quantitative results showing that our proposed method outperforms the considered baselines on downstream tasks, such as video prediction and visual question-answering tasks.

☆ Students Rather Than Experts: A New AI For Education Pipeline To Model More Human-Like And Personalised Early Adolescences

Yiping Ma, Shiyu Hu, Xuchen Li, Yipei Wang, Shiqing Liu, Kang Hao Cheong

The capabilities of large language models (LLMs) have been applied in expert systems across various domains, providing new opportunities for AI in Education. Educational interactions involve a cyclical exchange between teachers and students. Current research predominantly focuses on using LLMs to simulate teachers, leveraging their expertise to enhance student learning outcomes. However, the simulation of students, which could improve teachers' instructional skills, has received insufficient attention due to the challenges of modeling and evaluating virtual students. This research asks: Can LLMs be utilized to develop virtual student agents that mimic human-like behavior and individual variability? Unlike expert systems focusing on knowledge delivery, virtual students must replicate learning difficulties, emotional responses, and linguistic uncertainties. These traits present significant challenges in both modeling and evaluation. To address these issues, this study focuses on language learning as a context for modeling virtual student agents. We propose a novel AI4Education framework, called SOE (Scene-Object-Evaluation), to systematically construct LVSA (LLM-based Virtual Student Agents). By curating a dataset of personalized teacher-student interactions with various personality traits, question types, and learning stages, and fine-tuning LLMs using LoRA, we conduct multi-dimensional evaluation experiments. Specifically, we: (1) develop a theoretical framework for generating LVSA; (2) integrate human subjective evaluation metrics into GPT-4 assessments, demonstrating a strong correlation between human evaluators and GPT-4 in judging LVSA authenticity; and (3) validate that LLMs can generate human-like, personalized virtual student agents in educational contexts, laying a foundation for future applications in pre-service teacher training and multi-agent simulation environments.

☆ PALMS: Plane-based Accessible Indoor Localization Using Mobile Smartphones

Yunqian Cheng, Roberto Manduchi

In this paper, we present PALMS, an innovative indoor global localization and relocalization system for mobile smartphones that utilizes publicly available floor plans. Unlike most vision-based methods that require constant visual input, our system adopts a dynamic form of localization that considers a single instantaneous observation and odometry data. The core contribution of this work is the introduction of a particle filter initialization method that leverages the Certainly Empty Space (CES) constraint along with principal orientation matching. This approach creates a spatial probability distribution of the device's location, significantly improving localization accuracy and reducing particle filter convergence time. Our experimental evaluations demonstrate that PALMS outperforms traditional methods with uniformly initialized particle filters, providing a more efficient and accessible approach to indoor wayfinding. By eliminating the need for prior environmental fingerprinting, PALMS provides a scalable and practical approach to indoor navigation.

comment: 7 pages, 3 figures, accepted to the 14th International Conference on Indoor Positioning and Indoor Navigation (IPIN) 2024, Best Presentation Award

☆ Enhancing SNN-based Spatio-Temporal Learning: A Benchmark Dataset and Cross-Modality Attention Model

Shibo Zhou, Bo Yang, Mengwen Yuan, Runhao Jiang, Rui Yan, Gang Pan, Huajin Tang

Spiking Neural Networks (SNNs), renowned for their low power consumption, brain-inspired architecture, and spatio-temporal representation capabilities, have garnered considerable attention in recent years. Similar to Artificial Neural Networks (ANNs), high-quality benchmark datasets are of great importance to the advances of SNNs. However, our analysis indicates that many prevalent neuromorphic datasets lack strong temporal correlation, preventing SNNs from fully exploiting their spatio-temporal representation capabilities. Meanwhile, the integration of event and frame modalities offers more comprehensive visual spatio-temporal information. Yet, the SNN-based cross-modality fusion remains underexplored. In this work, we present a neuromorphic dataset called DVS-SLR that can better exploit the inherent spatio-temporal properties of SNNs. Compared to existing datasets, it offers advantages in terms of higher temporal correlation, larger scale, and more varied scenarios. In addition, our neuromorphic dataset contains corresponding frame data, which can be used for developing SNN-based fusion methods. By virtue of the dual-modal feature of the dataset, we propose a Cross-Modality Attention (CMA) based fusion method. The CMA model efficiently utilizes the unique advantages of each modality, allowing for SNNs to learn both temporal and spatial attention scores from the spatio-temporal features of event and frame modalities, subsequently allocating these scores across modalities to enhance their synergy. Experimental results demonstrate that our method not only improves recognition accuracy but also ensures robustness across diverse scenarios.

☆ RANSAC Back to SOTA: A Two-stage Consensus Filtering for Real-time 3D Registration

Pengcheng Shi, Shaocheng Yan, Yilin Xiao, Xinyi Liu, Yongjun Zhang, Jiayuan Li

Correspondence-based point cloud registration (PCR) plays a key role in robotics and computer vision. However, challenges like sensor noises, object occlusions, and descriptor limitations inevitably result in numerous outliers. RANSAC family is the most popular outlier removal solution. However, the requisite iterations escalate exponentially with the outlier ratio, rendering it far inferior to existing methods (SC2PCR [1], MAC [2], etc.) in terms of accuracy or speed. Thus, we propose a two-stage consensus filtering (TCF) that elevates RANSAC to state-of-the-art (SOTA) speed and accuracy. Firstly, one-point RANSAC obtains a consensus set based on length consistency. Subsequently, two-point RANSAC refines the set via angle consistency. Then, three-point RANSAC computes a coarse pose and removes outliers based on transformed correspondence's distances. Drawing on optimizations from one-point and two-point RANSAC, three-point RANSAC requires only a few iterations. Eventually, an iterative reweighted least squares (IRLS) is applied to yield the optimal pose. Experiments on the large-scale KITTI and ETH datasets demonstrate our method achieves up to three-orders-of-magnitude speedup compared to MAC while maintaining registration accuracy and recall. Our code is available at https://github.com/ShiPC-AI/TCF.

comment: 8 pages, 8 figures

☆ TALoS: Enhancing Semantic Scene Completion via Test-time Adaptation on the Line of Sight NeurIPS 2024

Hyun-Kurl Jang, Jihun Kim, Hyeokjun Kweon, Kuk-Jin Yoon

Semantic Scene Completion (SSC) aims to perform geometric completion and semantic segmentation simultaneously. Despite the promising results achieved by existing studies, the inherently ill-posed nature of the task presents significant challenges in diverse driving scenarios. This paper introduces TALoS, a novel test-time adaptation approach for SSC that excavates the information available in driving environments. Specifically, we focus on that observations made at a certain moment can serve as Ground Truth (GT) for scene completion at another moment. Given the characteristics of the LiDAR sensor, an observation of an object at a certain location confirms both 1) the occupation of that location and 2) the absence of obstacles along the line of sight from the LiDAR to that point. TALoS utilizes these observations to obtain self-supervision about occupancy and emptiness, guiding the model to adapt to the scene in test time. In a similar manner, we aggregate reliable SSC predictions among multiple moments and leverage them as semantic pseudo-GT for adaptation. Further, to leverage future observations that are not accessible at the current time, we present a dual optimization scheme using the model in which the update is delayed until the future observation is available. Evaluations on the SemanticKITTI validation and test sets demonstrate that TALoS significantly improves the performance of the pre-trained SSC model. Our code is available at https://github.com/blue-531/TALoS.

comment: Accepted at NeurIPS 2024. Code is available at https://github.com/blue-531/TALoS

☆ Transforming Blood Cell Detection and Classification with Advanced Deep Learning Models: A Comparative Study

Shilpa Choudhary, Sandeep Kumar, Pammi Sri Siddhaarth, Guntu Charitasri

Efficient detection and classification of blood cells are vital for accurate diagnosis and effective treatment of blood disorders. This study utilizes a YOLOv10 model trained on Roboflow data with images resized to 640x640 pixels across varying epochs. The results show that increased training epochs significantly enhance accuracy, precision, and recall, particularly in real-time blood cell detection & classification. The YOLOv10 model outperforms MobileNetV2, ShuffleNetV2, and DarkNet in real-time performance, though MobileNetV2 and ShuffleNetV2 are more computationally efficient, and DarkNet excels in feature extraction for blood cell classification. This research highlights the potential of integrating deep learning models like YOLOv10, MobileNetV2, ShuffleNetV2, and DarkNet into clinical workflows, promising improvements in diagnostic accuracy and efficiency. Additionally, a new, well-annotated blood cell dataset was created and will be open-sourced to support further advancements in automatic blood cell detection and classification. The findings demonstrate the transformative impact of these models in revolutionizing medical diagnostics and enhancing blood disorder management

comment: 26 pages, 4884 Words, 17 Figures, 10 Tables

☆ Calibration of ordinal regression networks

Daehwan Kim, Haejun Chung, Ikbeom Jang

Recent studies have shown that deep neural networks are not well-calibrated and produce over-confident predictions. The miscalibration issue primarily stems from the minimization of cross-entropy, which aims to align predicted softmax probabilities with one-hot labels. In ordinal regression tasks, this problem is compounded by an additional challenge: the expectation that softmax probabilities should exhibit unimodal distribution is not met with cross-entropy. Rather, the ordinal regression literature has focused on unimodality and overlooked calibration. To address these issues, we propose a novel loss function that introduces order-aware calibration, ensuring that prediction confidence adheres to ordinal relationships between classes. It incorporates soft ordinal encoding and label-smoothing-based regularization to enforce both calibration and unimodality. Extensive experiments across three popular ordinal regression benchmarks demonstrate that our approach achieves state-of-the-art calibration without compromising accuracy.

☆ CL-HOI: Cross-Level Human-Object Interaction Distillation from Vision Large Language Models

Jianjun Gao, Chen Cai, Ruoyu Wang, Wenyang Liu, Kim-Hui Yap, Kratika Garg, Boon-Siew Han

Human-object interaction (HOI) detection has seen advancements with Vision Language Models (VLMs), but these methods often depend on extensive manual annotations. Vision Large Language Models (VLLMs) can inherently recognize and reason about interactions at the image level but are computationally heavy and not designed for instance-level HOI detection. To overcome these limitations, we propose a Cross-Level HOI distillation (CL-HOI) framework, which distills instance-level HOIs from VLLMs image-level understanding without the need for manual annotations. Our approach involves two stages: context distillation, where a Visual Linguistic Translator (VLT) converts visual information into linguistic form, and interaction distillation, where an Interaction Cognition Network (ICN) reasons about spatial, visual, and context relations. We design contrastive distillation losses to transfer image-level context and interaction knowledge from the teacher to the student model, enabling instance-level HOI detection. Evaluations on HICO-DET and V-COCO datasets demonstrate that our CL-HOI surpasses existing weakly supervised methods and VLLM supervised methods, showing its efficacy in detecting HOIs without manual labels.

☆ Resource-Efficient Medical Report Generation using Large Language Models

Abdullah, Ameer Hamza, Seong Tae Kim

Medical report generation is the task of automatically writing radiology reports for chest X-ray images. Manually composing these reports is a time-consuming process that is also prone to human errors. Generating medical reports can therefore help reduce the burden on radiologists. In other words, we can promote greater clinical automation in the medical domain. In this work, we propose a new framework leveraging vision-enabled Large Language Models (LLM) for the task of medical report generation. We introduce a lightweight solution that achieves better or comparative performance as compared to previous solutions on the task of medical report generation. We conduct extensive experiments exploring different model sizes and enhancement approaches, such as prefix tuning to improve the text generation abilities of the LLMs. We evaluate our approach on a prominent large-scale radiology report dataset - MIMIC-CXR. Our results demonstrate the capability of our resource-efficient framework to generate patient-specific reports with strong medical contextual understanding and high precision.

☆ LucidFusion: Generating 3D Gaussians with Arbitrary Unposed Images

Hao He, Yixun Liang, Luozhou Wang, Yuanhao Cai, Xinli Xu, Hao-Xiang Guo, Xiang Wen, Yingcong Chen

Recent large reconstruction models have made notable progress in generating high-quality 3D objects from single images. However, these methods often struggle with controllability, as they lack information from multiple views, leading to incomplete or inconsistent 3D reconstructions. To address this limitation, we introduce LucidFusion, a flexible end-to-end feed-forward framework that leverages the Relative Coordinate Map (RCM). Unlike traditional methods linking images to 3D world thorough pose, LucidFusion utilizes RCM to align geometric features coherently across different views, making it highly adaptable for 3D generation from arbitrary, unposed images. Furthermore, LucidFusion seamlessly integrates with the original single-image-to-3D pipeline, producing detailed 3D Gaussians at a resolution of $512 \times 512$, making it well-suited for a wide range of applications.

comment: 17 pages, 12 figures, project page: coming soon

☆ Fully Explicit Dynamic Gaussian Splatting NeurIPS 2024

Junoh Lee, Chang-Yeon Won, Hyunjun Jung, Inhwan Bae, Hae-Gon Jeon

3D Gaussian Splatting has shown fast and high-quality rendering results in static scenes by leveraging dense 3D prior and explicit representations. Unfortunately, the benefits of the prior and representation do not involve novel view synthesis for dynamic motions. Ironically, this is because the main barrier is the reliance on them, which requires increasing training and rendering times to account for dynamic motions. In this paper, we design a Explicit 4D Gaussian Splatting(Ex4DGS). Our key idea is to firstly separate static and dynamic Gaussians during training, and to explicitly sample positions and rotations of the dynamic Gaussians at sparse timestamps. The sampled positions and rotations are then interpolated to represent both spatially and temporally continuous motions of objects in dynamic scenes as well as reducing computational cost. Additionally, we introduce a progressive training scheme and a point-backtracking technique that improves Ex4DGS's convergence. We initially train Ex4DGS using short timestamps and progressively extend timestamps, which makes it work well with a few point clouds. The point-backtracking is used to quantify the cumulative error of each Gaussian over time, enabling the detection and removal of erroneous Gaussians in dynamic scenes. Comprehensive experiments on various scenes demonstrate the state-of-the-art rendering quality from our method, achieving fast rendering of 62 fps on a single 2080Ti GPU.

comment: Accepted at NeurIPS 2024

☆ Towards Kriging-informed Conditional Diffusion for Regional Sea-Level Data Downscaling

Subhankar Ghosh, Arun Sharma, Jayant Gupta, Aneesh Subramanian, Shashi Shekhar

Given coarser-resolution projections from global climate models or satellite data, the downscaling problem aims to estimate finer-resolution regional climate data, capturing fine-scale spatial patterns and variability. Downscaling is any method to derive high-resolution data from low-resolution variables, often to provide more detailed and local predictions and analyses. This problem is societally crucial for effective adaptation, mitigation, and resilience against significant risks from climate change. The challenge arises from spatial heterogeneity and the need to recover finer-scale features while ensuring model generalization. Most downscaling methods \cite{Li2020} fail to capture the spatial dependencies at finer scales and underperform on real-world climate datasets, such as sea-level rise. We propose a novel Kriging-informed Conditional Diffusion Probabilistic Model (Ki-CDPM) to capture spatial variability while preserving fine-scale features. Experimental results on climate data show that our proposed method is more accurate than state-of-the-art downscaling techniques.

☆ Erasing Undesirable Concepts in Diffusion Models with Adversarial Preservation

Anh Bui, Long Vuong, Khanh Doan, Trung Le, Paul Montague, Tamas Abraham, Dinh Phung

Diffusion models excel at generating visually striking content from text but can inadvertently produce undesirable or harmful content when trained on unfiltered internet data. A practical solution is to selectively removing target concepts from the model, but this may impact the remaining concepts. Prior approaches have tried to balance this by introducing a loss term to preserve neutral content or a regularization term to minimize changes in the model parameters, yet resolving this trade-off remains challenging. In this work, we propose to identify and preserving concepts most affected by parameter changes, termed as \textit{adversarial concepts}. This approach ensures stable erasure with minimal impact on the other concepts. We demonstrate the effectiveness of our method using the Stable Diffusion model, showing that it outperforms state-of-the-art erasure methods in eliminating unwanted content while maintaining the integrity of other unrelated elements. Our code is available at \url{https://github.com/tuananhbui89/Erasing-Adversarial-Preservation}.

☆ Joint Top-Down and Bottom-Up Frameworks for 3D Visual Grounding ICPR2024

Yang Liu, Daizong Liu, Wei Hu

This paper tackles the challenging task of 3D visual grounding-locating a specific object in a 3D point cloud scene based on text descriptions. Existing methods fall into two categories: top-down and bottom-up methods. Top-down methods rely on a pre-trained 3D detector to generate and select the best bounding box, resulting in time-consuming processes. Bottom-up methods directly regress object bounding boxes with coarse-grained features, producing worse results. To combine their strengths while addressing their limitations, we propose a joint top-down and bottom-up framework, aiming to enhance the performance while improving the efficiency. Specifically, in the first stage, we propose a bottom-up based proposal generation module, which utilizes lightweight neural layers to efficiently regress and cluster several coarse object proposals instead of using a complex 3D detector. Then, in the second stage, we introduce a top-down based proposal consolidation module, which utilizes graph design to effectively aggregate and propagate the query-related object contexts among the generated proposals for further refinement. By jointly training these two modules, we can avoid the inherent drawbacks of the complex proposals in the top-down framework and the coarse proposals in the bottom-up framework. Experimental results on the ScanRefer benchmark show that our framework is able to achieve the state-of-the-art performance.

comment: Accepted by ICPR2024

☆ Topology-Aware Exploration of Circle of Willis for CTA and MRA: Segmentation, Detection, and Classification MICCAI 2024

Minghui Zhang, Xin You, Hanxiao Zhang, Yun Gu

The Circle of Willis (CoW) vessels is critical to connecting major circulations of the brain. The topology of the vascular structure is clinical significance to evaluate the risk, severity of the neuro-vascular diseases. The CoW has two representative angiographic imaging modalities, computed tomography angiography (CTA) and magnetic resonance angiography (MRA). TopCow24 provided 125 paired CTA-MRA dataset for the analysis of CoW. To explore both CTA and MRA images in a unified framework to learn the inherent topology of Cow, we construct the universal dataset via independent intensity preprocess, followed by joint resampling and normarlization. Then, we utilize the topology-aware loss to enhance the topology completeness of the CoW and the discrimination between different classes. A complementary topology-aware refinement is further conducted to enhance the connectivity within the same class. Our method was evaluated on all the three tasks and two modalities, achieving competitive results. In the final test phase of TopCow24 Challenge, we achieved the second place in the CTA-Seg-Task, the third palce in the CTA-Box-Task, the first place in the CTA-Edg-Task, the second place in the MRA-Seg-Task, the third palce in the MRA-Box-Task, the second place in the MRA-Edg-Task.

comment: Participation technical report for TopCoW24 challenge @ MICCAI 2024

☆ Exploring Stronger Transformer Representation Learning for Occluded Person Re-Identificatio

Zhangjian Ji, Donglin Cheng, Kai Feng

Due to some complex factors (e.g., occlusion, pose variation and diverse camera perspectives), extracting stronger feature representation in person re-identification remains a challenging task. In this paper, we proposed a novel self-supervision and supervision combining transformer-based person re-identification framework, namely SSSC-TransReID. Different from the general transformer-based person re-identification models, we designed a self-supervised contrastive learning branch, which can enhance the feature representation for person re-identification without negative samples or additional pre-training. In order to train the contrastive learning branch, we also proposed a novel random rectangle mask strategy to simulate the occlusion in real scenes, so as to enhance the feature representation for occlusion. Finally, we utilized the joint-training loss function to integrate the advantages of supervised learning with ID tags and self-supervised contrastive learning without negative samples, which can reinforce the ability of our model to excavate stronger discriminative features, especially for occlusion. Extensive experimental results on several benchmark datasets show our proposed model obtains superior Re-ID performance consistently and outperforms the state-of-the-art ReID methods by large margins on the mean average accuracy (mAP) and Rank-1 accuracy.

☆ Deep Active Learning with Manifold-preserving Trajectory Sampling

Yingrui Ji, Vijaya Sindhoori Kaza, Nishanth Artham, Tianyang Wang

Active learning (AL) is for optimizing the selection of unlabeled data for annotation (labeling), aiming to enhance model performance while minimizing labeling effort. The key question in AL is which unlabeled data should be selected for annotation. Existing deep AL methods arguably suffer from bias incurred by clabeled data, which takes a much lower percentage than unlabeled data in AL context. We observe that such an issue is severe in different types of data, such as vision and non-vision data. To address this issue, we propose a novel method, namely Manifold-Preserving Trajectory Sampling (MPTS), aiming to enforce the feature space learned from labeled data to represent a more accurate manifold. By doing so, we expect to effectively correct the bias incurred by labeled data, which can cause a biased selection of unlabeled data. Despite its focus on manifold, the proposed method can be conveniently implemented by performing distribution mapping with MMD (Maximum Mean Discrepancies). Extensive experiments on various vision and non-vision benchmark datasets demonstrate the superiority of our method. Our source code can be found here.

☆ P-YOLOv8: Efficient and Accurate Real-Time Detection of Distracted Driving

Mohamed R. Elshamy, Heba M. Emara, Mohamed R. Shoaib, Abdel-Hameed A. Badawy

Distracted driving is a critical safety issue that leads to numerous fatalities and injuries worldwide. This study addresses the urgent need for efficient and real-time machine learning models to detect distracted driving behaviors. Leveraging the Pretrained YOLOv8 (P-YOLOv8) model, a real-time object detection system is introduced, optimized for both speed and accuracy. This approach addresses the computational constraints and latency limitations commonly associated with conventional detection models. The study demonstrates P-YOLOv8 versatility in both object detection and image classification tasks using the Distracted Driver Detection dataset from State Farm, which includes 22,424 images across ten behavior categories. Our research explores the application of P-YOLOv8 for image classification, evaluating its performance compared to deep learning models such as VGG16, VGG19, and ResNet. Some traditional models often struggle with low accuracy, while others achieve high accuracy but come with high computational costs and slow detection speeds, making them unsuitable for real-time applications. P-YOLOv8 addresses these issues by achieving competitive accuracy with significant computational cost and efficiency advantages. In particular, P-YOLOv8 generates a lightweight model with a size of only 2.84 MB and a lower number of parameters, totaling 1,451,098, due to its innovative architecture. It achieves a high accuracy of 99.46 percent with this small model size, opening new directions for deployment on inexpensive and small embedded devices using Tiny Machine Learning (TinyML). The experimental results show robust performance, making P-YOLOv8 a cost-effective solution for real-time deployment. This study provides a detailed analysis of P-YOLOv8's architecture, training, and performance benchmarks, highlighting its potential for real-time use in detecting distracted driving.

★ Deep Learning and Machine Learning -- Object Detection and Semantic Segmentation: From Theory to Applications

Jintao Ren, Ziqian Bi, Qian Niu, Junyu Liu, Benji Peng, Sen Zhang, Xuanhe Pan, Jinlang Wang, Keyu Chen, Caitlyn Heqi Yin, Pohsun Feng, Yizhu Wen, Tianyang Wang, Silin Chen, Ming Li, Jiawei Xu, Ming Liu

This book offers an in-depth exploration of object detection and semantic segmentation, combining theoretical foundations with practical applications. It covers state-of-the-art advancements in machine learning and deep learning, with a focus on convolutional neural networks (CNNs), YOLO architectures, and transformer-based approaches like DETR. The book also delves into the integration of artificial intelligence (AI) techniques and large language models for enhanced object detection in complex environments. A thorough discussion of big data analysis is presented, highlighting the importance of data processing, model optimization, and performance evaluation metrics. By bridging the gap between traditional methods and modern deep learning frameworks, this book serves as a comprehensive guide for researchers, data scientists, and engineers aiming to leverage AI-driven methodologies in large-scale object detection tasks.

comment: 167 pages

☆ ARTS: Semi-Analytical Regressor using Disentangled Skeletal Representations for Human Mesh Recovery from Videos ACM MM 2024

Tao Tang, Hong Liu, Yingxuan You, Ti Wang, Wenhao Li

Although existing video-based 3D human mesh recovery methods have made significant progress, simultaneously estimating human pose and shape from low-resolution image features limits their performance. These image features lack sufficient spatial information about the human body and contain various noises (e.g., background, lighting, and clothing), which often results in inaccurate pose and inconsistent motion. Inspired by the rapid advance in human pose estimation, we discover that compared to image features, skeletons inherently contain accurate human pose and motion. Therefore, we propose a novel semiAnalytical Regressor using disenTangled Skeletal representations for human mesh recovery from videos, called ARTS. Specifically, a skeleton estimation and disentanglement module is proposed to estimate the 3D skeletons from a video and decouple them into disentangled skeletal representations (i.e., joint position, bone length, and human motion). Then, to fully utilize these representations, we introduce a semi-analytical regressor to estimate the parameters of the human mesh model. The regressor consists of three modules: Temporal Inverse Kinematics (TIK), Bone-guided Shape Fitting (BSF), and Motion-Centric Refinement (MCR). TIK utilizes joint position to estimate initial pose parameters and BSF leverages bone length to regress bone-aligned shape parameters. Finally, MCR combines human motion representation with image features to refine the initial human model parameters. Extensive experiments demonstrate that our ARTS surpasses existing state-of-the-art video-based methods in both per-frame accuracy and temporal consistency on popular benchmarks: 3DPW, MPI-INF-3DHP, and Human3.6M. Code is available at https://github.com/TangTao-PKU/ARTS.

comment: Accepted by ACM MM 2024. Project page: https://github.com/TangTao-PKU/ARTS

☆ Multimodal Learning for Embryo Viability Prediction in Clinical IVF MICCAI 2024

Junsik Kim, Zhiyi Shi, Davin Jeong, Johannes Knittel, Helen Y. Yang, Yonghyun Song, Wanhua Li, Yicong Li, Dalit Ben-Yosef, Daniel Needleman, Hanspeter Pfister

In clinical In-Vitro Fertilization (IVF), identifying the most viable embryo for transfer is important to increasing the likelihood of a successful pregnancy. Traditionally, this process involves embryologists manually assessing embryos' static morphological features at specific intervals using light microscopy. This manual evaluation is not only time-intensive and costly, due to the need for expert analysis, but also inherently subjective, leading to variability in the selection process. To address these challenges, we develop a multimodal model that leverages both time-lapse video data and Electronic Health Records (EHRs) to predict embryo viability. One of the primary challenges of our research is to effectively combine time-lapse video and EHR data, owing to their inherent differences in modality. We comprehensively analyze our multimodal model with various modality inputs and integration approaches. Our approach will enable fast and automated embryo viability predictions in scale for clinical IVF.

comment: Accepted to MICCAI 2024

☆ Online Pseudo-Label Unified Object Detection for Multiple Datasets Training

XiaoJun Tang, Jingru Wang, Zeyu Shangguan, Darun Tang, Yuyu Liu

The Unified Object Detection (UOD) task aims to achieve object detection of all merged categories through training on multiple datasets, and is of great significance in comprehensive object detection scenarios. In this paper, we conduct a thorough analysis of the cross datasets missing annotations issue, and propose an Online Pseudo-Label Unified Object Detection scheme. Our method uses a periodically updated teacher model to generate pseudo-labels for the unlabelled objects in each sub-dataset. This periodical update strategy could better ensure that the accuracy of the teacher model reaches the local maxima and maximized the quality of pseudo-labels. In addition, we survey the influence of overlapped region proposals on the accuracy of box regression. We propose a category specific box regression and a pseudo-label RPN head to improve the recall rate of the Region Proposal Network (PRN). Our experimental results on common used benchmarks (\eg COCO, Object365 and OpenImages) indicates that our online pseudo-label UOD method achieves higher accuracy than existing SOTA methods.

☆ A Dual Process VLA: Efficient Robotic Manipulation Leveraging VLM

ByungOk Han, Jaehong Kim, Jinhyeok Jang

Vision-Language-Action (VLA) models are receiving increasing attention for their ability to enable robots to perform complex tasks by integrating visual context with linguistic commands. However, achieving efficient real-time performance remains challenging due to the high computational demands of existing models. To overcome this, we propose Dual Process VLA (DP-VLA), a hierarchical framework inspired by dual-process theory. DP-VLA utilizes a Large System 2 Model (L-Sys2) for complex reasoning and decision-making, while a Small System 1 Model (S-Sys1) handles real-time motor control and sensory processing. By leveraging Vision-Language Models (VLMs), the L-Sys2 operates at low frequencies, reducing computational overhead, while the S-Sys1 ensures fast and accurate task execution. Experimental results on the RoboCasa dataset demonstrate that DP-VLA achieves faster inference and higher task success rates, providing a scalable solution for advanced robotic applications.

comment: 10 page

♻ ☆ Toward Generalizing Visual Brain Decoding to Unseen Subjects

Xiangtao Kong, Kexin Huang, Ping Li, Lei Zhang

Visual brain decoding aims to decode visual information from human brain activities. Despite the great progress, one critical limitation of current brain decoding research lies in the lack of generalization capability to unseen subjects. Prior works typically focus on decoding brain activity of individuals based on the observation that different subjects exhibit different brain activities, while it remains unclear whether brain decoding can be generalized to unseen subjects. This study aims to answer this question. We first consolidate an image-fMRI dataset consisting of stimulus-image and fMRI-response pairs, involving 177 subjects in the movie-viewing task of the Human Connectome Project (HCP). This dataset allows us to investigate the brain decoding performance with the increase of participants. We then present a learning paradigm that applies uniform processing across all subjects, instead of employing different network heads or tokenizers for individuals as in previous methods, which can accommodate a large number of subjects to explore the generalization capability across different subjects. A series of experiments are conducted and we have the following findings. First, the network exhibits clear generalization capabilities with the increase of training subjects. Second, the generalization capability is common to popular network architectures (MLP, CNN and Transformer). Third, the generalization performance is affected by the similarity between subjects. Our findings reveal the inherent similarities in brain activities across individuals. With the emerging of larger and more comprehensive datasets, it is possible to train a brain decoding foundation model in the future. Codes and models can be found at https://github.com/Xiangtaokong/TGBD.

♻ ☆ Utilizing Large Language Models in An Iterative Paradigm with Domain Feedback for Molecule Optimization

Khiem Le, Nitesh V. Chawla

Molecule optimization is a critical task in drug discovery to optimize desired properties of a given molecule through chemical modification. Despite Large Language Models (LLMs) holding the potential to efficiently simulate this task by using natural language to direct the optimization, straightforwardly utilizing shows limited performance. In this work, we facilitate utilizing LLMs in an iterative paradigm by proposing a simple yet highly effective domain feedback provider, namely $\text{Re}^2$DF. In detail, $\text{Re}^2$DF harnesses an external toolkit, RDKit, to handle the molecule hallucination, if the modified molecule is chemically invalid. Otherwise, its desired properties are computed and compared to the original one, establishing reliable domain feedback with correct direction and distance towards the objective, followed by a retrieved example, to explicitly guide the LLM to refine the modified molecule. We conduct experiments across both single- and multi-property objectives with 2 thresholds, where $\text{Re}^2$DF shows significant improvements. Particularly, for 20 single-property objectives, $\text{Re}^2$DF enhances Hit ratio by 16.95% and 20.76% under loose and strict thresholds, respectively. For 32 multi-property objectives, $\text{Re}^2$DF enhances Hit ratio by 6.04% and 5.25%.

♻ ☆ Decomposing and Interpreting Image Representations via Text in ViTs Beyond CLIP NeurIPS 2024

Sriram Balasubramanian, Samyadeep Basu, Soheil Feizi

Recent work has explored how individual components of the CLIP-ViT model contribute to the final representation by leveraging the shared image-text representation space of CLIP. These components, such as attention heads and MLPs, have been shown to capture distinct image features like shape, color or texture. However, understanding the role of these components in arbitrary vision transformers (ViTs) is challenging. To this end, we introduce a general framework which can identify the roles of various components in ViTs beyond CLIP. Specifically, we (a) automate the decomposition of the final representation into contributions from different model components, and (b) linearly map these contributions to CLIP space to interpret them via text. Additionally, we introduce a novel scoring function to rank components by their importance with respect to specific features. Applying our framework to various ViT variants (e.g. DeiT, DINO, DINOv2, Swin, MaxViT), we gain insights into the roles of different components concerning particular image features. These insights facilitate applications such as image retrieval using text descriptions or reference images, visualizing token importance heatmaps, and mitigating spurious correlations. We release our code to reproduce the experiments at https://github.com/SriramB-98/vit-decompose

comment: NeurIPS 2024, 31 pages, 15 figures

♻ ☆ RACCooN: A Versatile Instructional Video Editing Framework with Auto-Generated Narratives

Jaehong Yoon, Shoubin Yu, Mohit Bansal

Recent video generative models primarily rely on carefully written text prompts for specific tasks, like inpainting or style editing. They require labor-intensive textual descriptions for input videos, hindering their flexibility to adapt personal/raw videos to user specifications. This paper proposes RACCooN, a versatile and user-friendly video-to-paragraph-to-video generative framework that supports multiple video editing capabilities such as removal, addition, and modification, through a unified pipeline. RACCooN consists of two principal stages: Video-to-Paragraph (V2P) and Paragraph-to-Video (P2V). In the V2P stage, we automatically describe video scenes in well-structured natural language, capturing both the holistic context and focused object details. Subsequently, in the P2V stage, users can optionally refine these descriptions to guide the video diffusion model, enabling various modifications to the input video, such as removing, changing subjects, and/or adding new objects. The proposed approach stands out from other methods through several significant contributions: (1) RACCooN suggests a multi-granular spatiotemporal pooling strategy to generate well-structured video descriptions, capturing both the broad context and object details without requiring complex human annotations, simplifying precise video content editing based on text for users. (2) Our video generative model incorporates auto-generated narratives or instructions to enhance the quality and accuracy of the generated content. (3) RACCooN also plans to imagine new objects in a given video, so users simply prompt the model to receive a detailed video editing plan for complex video editing. The proposed framework demonstrates impressive versatile capabilities in video-to-paragraph generation, video content editing, and can be incorporated into other SoTA video generative models for further enhancement.

comment: The first two authors contribute equally. Project Page: https://raccoon-mllm-gen.github.io/

♻ ☆ Human-Agent Joint Learning for Efficient Robot Manipulation Skill Acquisition

Shengcheng Luo, Quanquan Peng, Jun Lv, Kaiwen Hong, Katherine Rose Driggs-Campbell, Cewu Lu, Yong-Lu Li

Employing a teleoperation system for gathering demonstrations offers the potential for more efficient learning of robot manipulation. However, teleoperating a robot arm equipped with a dexterous hand or gripper, via a teleoperation system presents inherent challenges due to the task's high dimensionality, complexity of motion, and differences between physiological structures. In this study, we introduce a novel system for joint learning between human operators and robots, that enables human operators to share control of a robot end-effector with a learned assistive agent, simplifies the data collection process, and facilitates simultaneous human demonstration collection and robot manipulation training. As data accumulates, the assistive agent gradually learns. Consequently, less human effort and attention are required, enhancing the efficiency of the data collection process. It also allows the human operator to adjust the control ratio to achieve a trade-off between manual and automated control. We conducted experiments in both simulated environments and physical real-world settings. Through user studies and quantitative evaluations, it is evident that the proposed system could enhance data collection efficiency and reduce the need for human adaptation while ensuring the collected data is of sufficient quality for downstream tasks. \textit{For more details, please refer to our webpage https://norweig1an.github.io/HAJL.github.io/.

comment: 8 pages, 6 figures

♻ ☆ CoTCoNet: An Optimized Coupled Transformer-Convolutional Network with an Adaptive Graph Reconstruction for Leukemia Detection

Chandravardhan Singh Raghaw, Arnav Sharma, Shubhi Bansal, Mohammad Zia Ur Rehman, Nagendra Kumar

Swift and accurate blood smear analysis is an effective diagnostic method for leukemia and other hematological malignancies. However, manual leukocyte count and morphological evaluation using a microscope is time-consuming and prone to errors. Conventional image processing methods also exhibit limitations in differentiating cells due to the visual similarity between malignant and benign cell morphology. This limitation is further compounded by the skewed training data that hinders the extraction of reliable and pertinent features. In response to these challenges, we propose an optimized Coupled Transformer Convolutional Network (CoTCoNet) framework for the classification of leukemia, which employs a well-designed transformer integrated with a deep convolutional network to effectively capture comprehensive global features and scalable spatial patterns, enabling the identification of complex and large-scale hematological features. Further, the framework incorporates a graph-based feature reconstruction module to reveal the hidden or unobserved hard-to-see biological features of leukocyte cells and employs a Population-based Meta-Heuristic Algorithm for feature selection and optimization. To mitigate data imbalance issues, we employ a synthetic leukocyte generator. In the evaluation phase, we initially assess CoTCoNet on a dataset containing 16,982 annotated cells, and it achieves remarkable accuracy and F1-Score rates of 0.9894 and 0.9893, respectively. To broaden the generalizability of our model, we evaluate it across four publicly available diverse datasets, which include the aforementioned dataset. This evaluation demonstrates that our method outperforms current state-of-the-art approaches. We also incorporate an explainability approach in the form of feature visualization closely aligned with cell annotations to provide a deeper understanding of the framework.

♻ ☆ PUMA: Empowering Unified MLLM with Multi-granular Visual Generation

Rongyao Fang, Chengqi Duan, Kun Wang, Hao Li, Hao Tian, Xingyu Zeng, Rui Zhao, Jifeng Dai, Hongsheng Li, Xihui Liu

Recent advancements in multimodal foundation models have yielded significant progress in vision-language understanding. Initial attempts have also explored the potential of multimodal large language models (MLLMs) for visual content generation. However, existing works have insufficiently addressed the varying granularity demands of different image generation tasks within a unified MLLM paradigm - from the diversity required in text-to-image generation to the precise controllability needed in image manipulation. In this work, we propose PUMA, emPowering Unified MLLM with Multi-grAnular visual generation. PUMA unifies multi-granular visual features as both inputs and outputs of MLLMs, elegantly addressing the different granularity requirements of various image generation tasks within a unified MLLM framework. Following multimodal pretraining and task-specific instruction tuning, PUMA demonstrates proficiency in a wide range of multimodal tasks. This work represents a significant step towards a truly unified MLLM capable of adapting to the granularity demands of various visual tasks. The code and model will be released in https://github.com/rongyaofang/PUMA.

comment: Project page: https://rongyaofang.github.io/puma/

♻ ☆ Pre-processing and Compression: Understanding Hidden Representation Refinement Across Imaging Domains via Intrinsic Dimension NeurIPS 2024

Nicholas Konz, Maciej A. Mazurowski

In recent years, there has been interest in how geometric properties such as intrinsic dimension (ID) of a neural network's hidden representations change through its layers, and how such properties are predictive of important model behavior such as generalization ability. However, evidence has begun to emerge that such behavior can change significantly depending on the domain of the network's training data, such as natural versus medical images. Here, we further this inquiry by exploring how the ID of a network's learned representations changes through its layers, in essence, characterizing how the network successively refines the information content of input data to be used for predictions. Analyzing eleven natural and medical image datasets across six network architectures, we find that how ID changes through the network differs noticeably between natural and medical image models. Specifically, medical image models peak in representation ID earlier in the network, implying a difference in the image features and their abstractness that are typically used for downstream tasks in these domains. Additionally, we discover a strong correlation of this peak representation ID with the ID of the data in its input space, implying that the intrinsic information content of a model's learned representations is guided by that of the data it was trained on. Overall, our findings emphasize notable discrepancies in network behavior between natural and non-natural imaging domains regarding hidden representation information content, and provide further insights into how a network's learned features are shaped by its training data.

comment: Published in NeurIPS 2024 Workshop on Scientific Methods for Understanding Deep Learning (SciForDL)

♻ ☆ SETA: Semantic-Aware Token Augmentation for Domain Generalization

Jintao Guo, Lei Qi, Yinghuan Shi, Yang Gao

Domain generalization (DG) aims to enhance the model robustness against domain shifts without accessing target domains. A prevalent category of methods for DG is data augmentation, which focuses on generating virtual samples to simulate domain shifts. However, existing augmentation techniques in DG are mainly tailored for convolutional neural networks (CNNs), with limited exploration in token-based architectures, i.e., vision transformer (ViT) and multi-layer perceptrons (MLP) models. In this paper, we study the impact of prior CNN-based augmentation methods on token-based models, revealing their performance is suboptimal due to the lack of incentivizing the model to learn holistic shape information. To tackle the issue, we propose the SEmantic-aware Token Augmentation (SETA) method. SETA transforms token features by perturbing local edge cues while preserving global shape features, thereby enhancing the model learning of shape information. To further enhance the generalization ability of the model, we introduce two stylized variants of our method combined with two state-of-the-art style augmentation methods in DG. We provide a theoretical insight into our method, demonstrating its effectiveness in reducing the generalization risk bound. Comprehensive experiments on five benchmarks prove that our method achieves SOTA performances across various ViT and MLP architectures. Our code is available at https://github.com/lingeringlight/SETA.

comment: Accepted by IEEE TIP 2024. The code is available at https://github.com/lingeringlight/SETA

♻ ☆ Machine Unlearning in Forgettability Sequence

Junjie Chen, Qian Chen, Jian Lou, Xiaoyu Zhang, Kai Wu, Zilong Wang

Machine unlearning (MU) is becoming a promising paradigm to achieve the "right to be forgotten", where the training trace of any chosen data points could be eliminated, while maintaining the model utility on general testing samples after unlearning. With the advancement of forgetting research, many fundamental open questions remain unanswered: do different samples exhibit varying levels of difficulty in being forgotten? Further, does the sequence in which samples are forgotten, determined by their respective difficulty levels, influence the performance of forgetting algorithms? In this paper, we identify key factor affecting unlearning difficulty and the performance of unlearning algorithms. We find that samples with higher privacy risks are more likely to be unlearning, indicating that the unlearning difficulty varies among different samples which motives a more precise unlearning mode. Built upon this insight, we propose a general unlearning framework, dubbed RSU, which consists of Ranking module and SeqUnlearn module.

comment: The senior authors of the draft are not fully convinced that the novelty is significant enough for this submission compared to the latest research progress in this area. Additionally, the senior authors have identified writing issues. Based on these two reasons, we have decided to withdraw the draft from arXiv

♻ ☆ From FDG to PSMA: A Hitchhiker's Guide to Multitracer, Multicenter Lesion Segmentation in PET/CT Imaging

Maximilian Rokuss, Balint Kovacs, Yannick Kirchhoff, Shuhan Xiao, Constantin Ulrich, Klaus H. Maier-Hein, Fabian Isensee

Automated lesion segmentation in PET/CT scans is crucial for improving clinical workflows and advancing cancer diagnostics. However, the task is challenging due to physiological variability, different tracers used in PET imaging, and diverse imaging protocols across medical centers. To address this, the autoPET series was created to challenge researchers to develop algorithms that generalize across diverse PET/CT environments. This paper presents our solution for the autoPET III challenge, targeting multitracer, multicenter generalization using the nnU-Net framework with the ResEncL architecture. Key techniques include misalignment data augmentation and multi-modal pretraining across CT, MR, and PET datasets to provide an initial anatomical understanding. We incorporate organ supervision as a multitask approach, enabling the model to distinguish between physiological uptake and tracer-specific patterns, which is particularly beneficial in cases where no lesions are present. Compared to the default nnU-Net, which achieved a Dice score of 57.61, or the larger ResEncL (65.31) our model significantly improved performance with a Dice score of 68.40, alongside a reduction in false positive (FPvol: 7.82) and false negative (FNvol: 10.35) volumes. These results underscore the effectiveness of combining advanced network design, augmentation, pretraining, and multitask learning for PET/CT lesion segmentation. After evaluation on the test set, our approach was awarded the first place in the model-centric category (Team LesionTracer). Code is publicly available at https://github.com/MIC-DKFZ/autopet-3-submission.

comment: Winning method of the autoPET III challenge (model-centric) - Team LesionTracer

♻ ☆ Deep Correlated Prompting for Visual Recognition with Missing Modalities NeurIPS 2024

Lianyu Hu, Tongkai Shi, Wei Feng, Fanhua Shang, Liang Wan

Large-scale multimodal models have shown excellent performance over a series of tasks powered by the large corpus of paired multimodal training data. Generally, they are always assumed to receive modality-complete inputs. However, this simple assumption may not always hold in the real world due to privacy constraints or collection difficulty, where models pretrained on modality-complete data easily demonstrate degraded performance on missing-modality cases. To handle this issue, we refer to prompt learning to adapt large pretrained multimodal models to handle missing-modality scenarios by regarding different missing cases as different types of input. Instead of only prepending independent prompts to the intermediate layers, we present to leverage the correlations between prompts and input features and excavate the relationships between different layers of prompts to carefully design the instructions. We also incorporate the complementary semantics of different modalities to guide the prompting design for each modality. Extensive experiments on three commonly-used datasets consistently demonstrate the superiority of our method compared to the previous approaches upon different missing scenarios. Plentiful ablations are further given to show the generalizability and reliability of our method upon different modality-missing ratios and types.

comment: NeurIPS 2024, add some results

♻ ☆ UNetMamba: An Efficient UNet-Like Mamba for Semantic Segmentation of High-Resolution Remote Sensing Images

Enze Zhu, Zhan Chen, Dingkai Wang, Hanru Shi, Xiaoxuan Liu, Lei Wang

Semantic segmentation of high-resolution remote sensing images is vital in downstream applications such as land-cover mapping, urban planning and disaster assessment.Existing Transformer-based methods suffer from the constraint between accuracy and efficiency, while the recently proposed Mamba is renowned for being efficient. Therefore, to overcome the dilemma, we propose UNetMamba, a UNet-like semantic segmentation model based on Mamba. It incorporates a mamba segmentation decoder (MSD) that can efficiently decode the complex information within high-resolution images, and a local supervision module (LSM), which is train-only but can significantly enhance the perception of local contents. Extensive experiments demonstrate that UNetMamba outperforms the state-of-the-art methods with mIoU increased by 0.87% on LoveDA and 0.39% on ISPRS Vaihingen, while achieving high efficiency through the lightweight design, less memory footprint and reduced computational cost. The source code is available at https://github.com/EnzeZhu2001/UNetMamba.

comment: 5 pages, 3 figures

♻ ☆ A gradient-based approach to fast and accurate head motion compensation in cone-beam CT

Mareike Thies, Fabian Wagner, Noah Maul, Haijun Yu, Manuela Goldmann, Linda-Sophie Schneider, Mingxuan Gu, Siyuan Mei, Lukas Folle, Alexander Preuhs, Michael Manhart, Andreas Maier

Cone-beam computed tomography (CBCT) systems, with their flexibility, present a promising avenue for direct point-of-care medical imaging, particularly in critical scenarios such as acute stroke assessment. However, the integration of CBCT into clinical workflows faces challenges, primarily linked to long scan duration resulting in patient motion during scanning and leading to image quality degradation in the reconstructed volumes. This paper introduces a novel approach to CBCT motion estimation using a gradient-based optimization algorithm, which leverages generalized derivatives of the backprojection operator for cone-beam CT geometries. Building on that, a fully differentiable target function is formulated which grades the quality of the current motion estimate in reconstruction space. We drastically accelerate motion estimation yielding a 19-fold speed-up compared to existing methods. Additionally, we investigate the architecture of networks used for quality metric regression and propose predicting voxel-wise quality maps, favoring autoencoder-like architectures over contracting ones. This modification improves gradient flow, leading to more accurate motion estimation. The presented method is evaluated through realistic experiments on head anatomy. It achieves a reduction in reprojection error from an initial average of 3mm to 0.61mm after motion compensation and consistently demonstrates superior performance compared to existing approaches. The analytic Jacobian for the backprojection operation, which is at the core of the proposed method, is made publicly available. In summary, this paper contributes to the advancement of CBCT integration into clinical workflows by proposing a robust motion estimation approach that enhances efficiency and accuracy, addressing critical challenges in time-sensitive scenarios.

comment: \copyright 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works

♻ ☆ VeLoRA: Memory Efficient Training using Rank-1 Sub-Token Projections NeurIPS 2024

Roy Miles, Pradyumna Reddy, Ismail Elezi, Jiankang Deng

Large language models (LLMs) have recently emerged as powerful tools for tackling many language-processing tasks. Despite their success, training and fine-tuning these models is still far too computationally and memory intensive. In this paper, we identify and characterise the important components needed for effective model convergence using gradient descent. In doing so we find that the intermediate activations used to implement backpropagation can be excessively compressed without incurring any degradation in performance. This result leads us to a cheap and memory-efficient algorithm for both fine-tuning and pre-training LLMs. The proposed algorithm simply divides the tokens up into smaller sub-tokens before projecting them onto a fixed 1-dimensional subspace during the forward pass. These features are then coarsely reconstructed during the backward pass to implement the update rules. We confirm the effectiveness of our algorithm as being complimentary to many state-of-the-art PEFT methods on the VTAB-1k fine-tuning benchmark. Furthermore, we outperform QLoRA for fine-tuning LLaMA and show competitive performance against other memory-efficient pre-training methods on the large-scale C4 dataset.

comment: NeurIPS 2024. Code available at https://github.com/roymiles/VeLoRA

♻ ☆ Towards Realistic Data Generation for Real-World Super-Resolution

Long Peng, Wenbo Li, Renjing Pei, Jingjing Ren, Yang Wang, Yang Cao, Zheng-Jun Zha

Existing image super-resolution (SR) techniques often fail to generalize effectively in complex real-world settings due to the significant divergence between training data and practical scenarios. To address this challenge, previous efforts have either manually simulated intricate physical-based degradations or utilized learning-based techniques, yet these approaches remain inadequate for producing large-scale, realistic, and diverse data simultaneously. In this paper, we introduce a novel Realistic Decoupled Data Generator (RealDGen), an unsupervised learning data generation framework designed for real-world super-resolution. We meticulously develop content and degradation extraction strategies, which are integrated into a novel content-degradation decoupled diffusion model to create realistic low-resolution images from unpaired real LR and HR images. Extensive experiments demonstrate that RealDGen excels in generating large-scale, high-quality paired data that mirrors real-world degradations, significantly advancing the performance of popular SR models on various real-world benchmarks.

♻ ★ CARLA Drone: Monocular 3D Object Detection from a Different Perspective

Johannes Meier, Luca Scalerandi, Oussema Dhaouadi, Jacques Kaiser, Nikita Araslanov, Daniel Cremers

Existing techniques for monocular 3D detection have a serious restriction. They tend to perform well only on a limited set of benchmarks, faring well either on ego-centric car views or on traffic camera views, but rarely on both. To encourage progress, this work advocates for an extended evaluation of 3D detection frameworks across different camera perspectives. We make two key contributions. First, we introduce the CARLA Drone dataset, CDrone. Simulating drone views, it substantially expands the diversity of camera perspectives in existing benchmarks. Despite its synthetic nature, CDrone represents a real-world challenge. To show this, we confirm that previous techniques struggle to perform well both on CDrone and a real-world 3D drone dataset. Second, we develop an effective data augmentation pipeline called GroundMix. Its distinguishing element is the use of the ground for creating 3D-consistent augmentation of a training image. GroundMix significantly boosts the detection accuracy of a lightweight one-stage detector. In our expanded evaluation, we achieve the average precision on par with or substantially higher than the previous state of the art across all tested datasets.

♻ ☆ UADA3D: Unsupervised Adversarial Domain Adaptation for 3D Object Detection with Sparse LiDAR and Large Domain Gaps

Maciej K Wozniak, Mattias Hansson, Marko Thiel, Patric Jensfelt

In this study, we address a gap in existing unsupervised domain adaptation approaches on LiDAR-based 3D object detection, which have predominantly concentrated on adapting between established, high-density autonomous driving datasets. We focus on sparser point clouds, capturing scenarios from different perspectives: not just from vehicles on the road but also from mobile robots on sidewalks, which encounter significantly different environmental conditions and sensor configurations. We introduce Unsupervised Adversarial Domain Adaptation for 3D Object Detection (UADA3D). UADA3D does not depend on pre-trained source models or teacher-student architectures. Instead, it uses an adversarial approach to directly learn domain-invariant features. We demonstrate its efficacy in various adaptation scenarios, showing significant improvements in both self-driving car and mobile robot domains. Our code is open-source and will be available soon.

comment: Accepted for IEEE RA-L 2024

♻ ☆ HeightFormer: A Semantic Alignment Monocular 3D Object Detection Method from Roadside Perspective

Pei Liu, Zihao Zhang, Haipeng Liu, Nanfang Zheng, Meixin Zhu, Ziyuan Pu

The on-board 3D object detection technology has received extensive attention as a critical technology for autonomous driving, while few studies have focused on applying roadside sensors in 3D traffic object detection. Existing studies achieve the projection of 2D image features to 3D features through height estimation based on the frustum. However, they did not consider the height alignment and the extraction efficiency of bird's-eye-view features. We propose a novel 3D object detection framework integrating Spatial Former and Voxel Pooling Former to enhance 2D-to-3D projection based on height estimation. Extensive experiments were conducted using the Rope3D and DAIR-V2X-I dataset, and the results demonstrated the outperformance of the proposed algorithm in the detection of both vehicles and cyclists. These results indicate that the algorithm is robust and generalized under various detection scenarios. Improving the accuracy of 3D object detection on the roadside is conducive to building a safe and trustworthy intelligent transportation system of vehicle-road coordination and promoting the large-scale application of autonomous driving. The code and pre-trained models will be released on https://anonymous.4open.science/r/HeightFormer.

♻ ☆ DARES: Depth Anything in Robotic Endoscopic Surgery with Self-supervised Vector-LoRA of the Foundation Model

Mona Sheikh Zeinoddin, Chiara Lena, Jiongqi Qu, Luca Carlini, Mattia Magro, Seunghoi Kim, Elena De Momi, Sophia Bano, Matthew Grech-Sollars, Evangelos Mazomenos, Daniel C. Alexander, Danail Stoyanov, Matthew J. Clarkson, Mobarakol Islam

Robotic-assisted surgery (RAS) relies on accurate depth estimation for 3D reconstruction and visualization. While foundation models like Depth Anything Models (DAM) show promise, directly applying them to surgery often yields suboptimal results. Fully fine-tuning on limited surgical data can cause overfitting and catastrophic forgetting, compromising model robustness and generalization. Although Low-Rank Adaptation (LoRA) addresses some adaptation issues, its uniform parameter distribution neglects the inherent feature hierarchy, where earlier layers, learning more general features, require more parameters than later ones. To tackle this issue, we introduce Depth Anything in Robotic Endoscopic Surgery (DARES), a novel approach that employs a new adaptation technique, Vector Low-Rank Adaptation (Vector-LoRA) on the DAM V2 to perform self-supervised monocular depth estimation in RAS scenes. To enhance learning efficiency, we introduce Vector-LoRA by integrating more parameters in earlier layers and gradually decreasing parameters in later layers. We also design a reprojection loss based on the multi-scale SSIM error to enhance depth perception by better tailoring the foundation model to the specific requirements of the surgical environment. The proposed method is validated on the SCARED dataset and demonstrates superior performance over recent state-of-the-art self-supervised monocular depth estimation techniques, achieving an improvement of 13.3% in the absolute relative error metric. The code and pre-trained weights are available at https://github.com/mobarakol/DARES.

comment: 11 pages

♻ ★ Any2Point: Empowering Any-modality Large Models for Efficient 3D Understanding

Yiwen Tang, Ray Zhang, Jiaming Liu, Zoey Guo, Dong Wang, Zhigang Wang, Bin Zhao, Shanghang Zhang, Peng Gao, Hongsheng Li, Xuelong Li

Large foundation models have recently emerged as a prominent focus of interest, attaining superior performance in widespread scenarios. Due to the scarcity of 3D data, many efforts have been made to adapt pre-trained transformers from vision to 3D domains. However, such 2D-to-3D approaches are still limited, due to the potential loss of spatial geometries and high computation cost. More importantly, their frameworks are mainly designed for 2D models, lacking a general any-to-3D paradigm. In this paper, we introduce Any2Point, a parameter-efficient method to empower any-modality large models (vision, language, audio) for 3D understanding. Given a frozen transformer from any source modality, we propose a 3D-to-any (1D or 2D) virtual projection strategy that correlates the input 3D points to the original 1D or 2D positions within the source modality. This mechanism enables us to assign each 3D token with a positional encoding paired with the pre-trained model, which avoids 3D geometry loss caused by the true projection and better motivates the transformer for 3D learning with 1D/2D positional priors. Then, within each transformer block, we insert an any-to-3D guided adapter module for parameter-efficient fine-tuning. The adapter incorporates prior spatial knowledge from the source modality to guide the local feature aggregation of 3D tokens, compelling the semantic adaption of any-modality transformers. We conduct extensive experiments to showcase the effectiveness and efficiency of our method. Code and models are released at https://github.com/Ivan-Tang-3D/Any2Point.

comment: Code and models are released at https://github.com/Ivan-Tang-3D/Any2Point

♻ ☆ Point-PEFT: Parameter-Efficient Fine-Tuning for 3D Pre-trained Models

Yiwen Tang, Ray Zhang, Zoey Guo, Dong Wang, Zhigang Wang, Bin Zhao, Xuelong Li

The popularity of pre-trained large models has revolutionized downstream tasks across diverse fields, such as language, vision, and multi-modality. To minimize the adaption cost for downstream tasks, many Parameter-Efficient Fine-Tuning (PEFT) techniques are proposed for language and 2D image pre-trained models. However, the specialized PEFT method for 3D pre-trained models is still under-explored. To this end, we introduce Point-PEFT, a novel framework for adapting point cloud pre-trained models with minimal learnable parameters. Specifically, for a pre-trained 3D model, we freeze most of its parameters, and only tune the newly added PEFT modules on downstream tasks, which consist of a Point-prior Prompt and a Geometry-aware Adapter. The Point-prior Prompt adopts a set of learnable prompt tokens, for which we propose to construct a memory bank with domain-specific knowledge, and utilize a parameter-free attention to enhance the prompt tokens. The Geometry-aware Adapter aims to aggregate point cloud features within spatial neighborhoods to capture fine-grained geometric information through local interactions. Extensive experiments indicate that our Point-PEFT can achieve better performance than the full fine-tuning on various downstream tasks, while using only 5% of the trainable parameters, demonstrating the efficiency and effectiveness of our approach. Code is released at https://github.com/Ivan-Tang-3D/Point-PEFT.

comment: The specialized PEFT framework for 3D pre-trained models, which achieves competitive performance to full fine-tuning, and significantly reduces the computational resources. Project page: https://github.com/Ivan-Tang-3D/Point-PEFT

♻ ☆ Fool Me Once? Contrasting Textual and Visual Explanations in a Clinical Decision-Support Setting EMNLP 2024

Maxime Kayser, Bayar Menzat, Cornelius Emde, Bogdan Bercean, Alex Novak, Abdala Espinosa, Bartlomiej W. Papiez, Susanne Gaube, Thomas Lukasiewicz, Oana-Maria Camburu

The growing capabilities of AI models are leading to their wider use, including in safety-critical domains. Explainable AI (XAI) aims to make these models safer to use by making their inference process more transparent. However, current explainability methods are seldom evaluated in the way they are intended to be used: by real-world end users. To address this, we conducted a large-scale user study with 85 healthcare practitioners in the context of human-AI collaborative chest X-ray analysis. We evaluated three types of explanations: visual explanations (saliency maps), natural language explanations, and a combination of both modalities. We specifically examined how different explanation types influence users depending on whether the AI advice and explanations are factually correct. We find that text-based explanations lead to significant over-reliance, which is alleviated by combining them with saliency maps. We also observe that the quality of explanations, that is, how much factually correct information they entail, and how much this aligns with AI correctness, significantly impacts the usefulness of the different explanation types.

comment: EMNLP 2024

♻ ☆ Diffusion Lens: Interpreting Text Encoders in Text-to-Image Pipelines ACL 2024

Michael Toker, Hadas Orgad, Mor Ventura, Dana Arad, Yonatan Belinkov

Text-to-image diffusion models (T2I) use a latent representation of a text prompt to guide the image generation process. However, the process by which the encoder produces the text representation is unknown. We propose the Diffusion Lens, a method for analyzing the text encoder of T2I models by generating images from its intermediate representations. Using the Diffusion Lens, we perform an extensive analysis of two recent T2I models. Exploring compound prompts, we find that complex scenes describing multiple objects are composed progressively and more slowly compared to simple scenes; Exploring knowledge retrieval, we find that representation of uncommon concepts requires further computation compared to common concepts, and that knowledge retrieval is gradual across layers. Overall, our findings provide valuable insights into the text encoder component in T2I pipelines.

comment: Published in: ACL 2024 Project webpage: tokeron.github.io/DiffusionLensWeb

♻ ☆ DriveDreamer4D: World Models Are Effective Data Machines for 4D Driving Scene Representation

Guosheng Zhao, Chaojun Ni, Xiaofeng Wang, Zheng Zhu, Xueyang Zhang, Yida Wang, Guan Huang, Xinze Chen, Boyuan Wang, Youyi Zhang, Wenjun Mei, Xingang Wang

Closed-loop simulation is essential for advancing end-to-end autonomous driving systems. Contemporary sensor simulation methods, such as NeRF and 3DGS, rely predominantly on conditions closely aligned with training data distributions, which are largely confined to forward-driving scenarios. Consequently, these methods face limitations when rendering complex maneuvers (e.g., lane change, acceleration, deceleration). Recent advancements in autonomous-driving world models have demonstrated the potential to generate diverse driving videos. However, these approaches remain constrained to 2D video generation, inherently lacking the spatiotemporal coherence required to capture intricacies of dynamic driving environments. In this paper, we introduce DriveDreamer4D, which enhances 4D driving scene representation leveraging world model priors. Specifically, we utilize the world model as a data machine to synthesize novel trajectory videos based on real-world driving data. Notably, we explicitly leverage structured conditions to control the spatial-temporal consistency of foreground and background elements, thus the generated data adheres closely to traffic constraints. To our knowledge, DriveDreamer4D is the first to utilize video generation models for improving 4D reconstruction in driving scenarios. Experimental results reveal that DriveDreamer4D significantly enhances generation quality under novel trajectory views, achieving a relative improvement in FID by 24.5%, 39.0%, and 10.5% compared to PVG, S3Gaussian, and Deformable-GS. Moreover, DriveDreamer4D markedly enhances the spatiotemporal coherence of driving agents, which is verified by a comprehensive user study and the relative increases of 20.3%, 42.0%, and 13.7% in the NTA-IoU metric.

comment: Project Page: https://drivedreamer4d.github.io

♻ ☆ Deep Multimodal Learning with Missing Modality: A Survey

Renjie Wu, Hu Wang, Hsiang-Ting Chen, Gustavo Carneiro

During multimodal model training and testing, certain data modalities may be absent due to sensor limitations, cost constraints, privacy concerns, or data loss, negatively affecting performance. Multimodal learning techniques designed to handle missing modalities can mitigate this by ensuring model robustness even when some modalities are unavailable. This survey reviews recent progress in Multimodal Learning with Missing Modality (MLMM), focusing on deep learning methods. It provides the first comprehensive survey that covers the motivation and distinctions between MLMM and standard multimodal learning setups, followed by a detailed analysis of current methods, applications, and datasets, concluding with challenges and future directions.

comment: Submitted to ACM Computing Surveys

♻ ☆ Shotluck Holmes: A Family of Efficient Small-Scale Large Language Vision Models For Video Captioning and Summarization

Richard Luo, Austin Peng, Adithya Vasudev, Rishabh Jain

Video is an increasingly prominent and information-dense medium, yet it poses substantial challenges for language models. A typical video consists of a sequence of shorter segments, or shots, that collectively form a coherent narrative. Each shot is analogous to a word in a sentence where multiple data streams of information (such as visual and auditory data) must be processed simultaneously. Comprehension of the entire video requires not only understanding the visual-audio information of each shot but also requires that the model links the ideas between each shot to generate a larger, all-encompassing story. Despite significant progress in the field, current works often overlook videos' more granular shot-by-shot semantic information. In this project, we propose a family of efficient large language vision models (LLVMs) to boost video summarization and captioning called Shotluck Holmes. By leveraging better pretraining and data collection strategies, we extend the abilities of existing small LLVMs from being able to understand a picture to being able to understand a sequence of frames. Specifically, we show that Shotluck Holmes achieves better performance than state-of-the-art results on the Shot2Story video captioning and summary task with significantly smaller and more computationally efficient models.

♻ ☆ LongVILA: Scaling Long-Context Visual Language Models for Long Videos

Fuzhao Xue, Yukang Chen, Dacheng Li, Qinghao Hu, Ligeng Zhu, Xiuyu Li, Yunhao Fang, Haotian Tang, Shang Yang, Zhijian Liu, Ethan He, Hongxu Yin, Pavlo Molchanov, Jan Kautz, Linxi Fan, Yuke Zhu, Yao Lu, Song Han

Long-context capability is critical for multi-modal foundation models, especially for long video understanding. We introduce LongVILA, a full-stack solution for long-context visual-language models \qinghao{by co-designing the algorithm and system. For model training, we upgrade existing VLMs to support long video understanding by incorporating two additional stages, {\em i.e.}, long context extension and long video supervised fine-tuning. However, training on long video is computationally and memory intensive. We introduce the long-context Multi-Modal Sequence Parallelism (MM-SP) system that efficiently parallelizes long video training and inference, enabling 2M context length training on 256 GPUs without any gradient checkpointing. LongVILA efficiently extends the number of video frames of VILA from 8 to 2048, improving the long video captioning score from 2.00 to 3.26 (out of 5), achieving 99.8% accuracy in 6,000-frame (more than 1 million tokens) video needle-in-a-haystack. LongVILA-7B demonstrates strong accuracy on the VideoMME benchmark, i.e., 61.8% with subtitle. Besides, MM-SP is 2.1x - 5.7x faster than ring style sequence parallelism and 1.1x - 1.4x faster than Megatron with a hybrid context and tensor parallelism. Moreover, it seamlessly integrates with Hugging Face Transformers.

comment: Code and models are available at https://github.com/NVlabs/VILA/blob/main/LongVILA.md

♻ ☆ Back-in-Time Diffusion: Unsupervised Detection of Medical Deepfakes

Fred Grabovski, Lior Yasur, Guy Amit, Yisroel Mirsky

Recent progress in generative models has made it easier for a wide audience to edit and create image content, raising concerns about the proliferation of deepfakes, especially in healthcare. Despite the availability of numerous techniques for detecting manipulated images captured by conventional cameras, their applicability to medical images is limited. This limitation stems from the distinctive forensic characteristics of medical images, a result of their imaging process. In this work we propose a novel anomaly detector for medical imagery based on diffusion models. Normally, diffusion models are used to generate images. However, we show how a similar process can be used to detect synthetic content by making a model reverse the diffusion on a suspected image. We evaluate our method on the task of detecting fake tumors injected and removed from CT and MRI scans. Our method significantly outperforms other state of the art unsupervised detectors with an increased AUC of 0.9 from 0.79 for injection and of 0.96 from 0.91 for removal on average. We also explore our hypothesis using AI explainability tools and publish our code and new medical deepfake datasets to encourage further research into this domain.

♻ ☆ Motion Segmentation for Neuromorphic Aerial Surveillance

Sami Arja, Alexandre Marcireau, Saeed Afshar, Bharath Ramesh, Gregory Cohen

Aerial surveillance demands rapid and precise detection of moving objects in dynamic environments. Event cameras, which draw inspiration from biological vision systems, present a promising alternative to frame-based sensors due to their exceptional temporal resolution, superior dynamic range, and minimal power requirements. Unlike traditional frame-based sensors that capture redundant information at fixed intervals, event cameras asynchronously record pixel-level brightness changes, providing a continuous and efficient data stream ideal for fast motion segmentation. While these sensors are ideal for fast motion segmentation, existing event-based motion segmentation methods often suffer from limitations such as the need for per-scene parameter tuning or reliance on manual labelling, hindering their scalability and practical deployment. In this paper, we address these challenges by introducing a novel motion segmentation method that leverages self-supervised vision transformers on both event data and optical flow information. Our approach eliminates the need for human annotations and reduces dependency on scene-specific parameters. In this paper, we used the EVK4-HD Prophesee event camera onboard a highly dynamic aerial platform in urban settings. We conduct extensive evaluations of our framework across multiple datasets, demonstrating state-of-the-art performance compared to existing benchmarks. Our method can effectively handle various types of motion and an arbitrary number of moving objects. Code and dataset are available at: \url{https://samiarja.github.io/evairborne/}

comment: 17 pages, 11 figures, 8 tables

♻ ☆ You Only Sample Once: Taming One-Step Text-to-Image Synthesis by Self-Cooperative Diffusion GANs

Yihong Luo, Xiaolong Chen, Xinghua Qu, Tianyang Hu, Jing Tang

Recently, some works have tried to combine diffusion and Generative Adversarial Networks (GANs) to alleviate the computational cost of the iterative denoising inference in Diffusion Models (DMs). However, existing works in this line suffer from either training instability and mode collapse or subpar one-step generation learning efficiency. To address these issues, we introduce YOSO, a novel generative model designed for rapid, scalable, and high-fidelity one-step image synthesis with high training stability and mode coverage. Specifically, we smooth the adversarial divergence by the denoising generator itself, performing self-cooperative learning. We show that our method can serve as a one-step generation model training from scratch with competitive performance. Moreover, we extend our YOSO to one-step text-to-image generation based on pre-trained models by several effective training techniques (i.e., latent perceptual loss and latent discriminator for efficient training along with the latent DMs; the informative prior initialization (IPI), and the quick adaption stage for fixing the flawed noise scheduler). Experimental results show that YOSO achieves the state-of-the-art one-step generation performance even with Low-Rank Adaptation (LoRA) fine-tuning. In particular, we show that the YOSO-PixArt-$\alpha$ can generate images in one step trained on 512 resolution, with the capability of adapting to 1024 resolution without extra explicit training, requiring only ~10 A800 days for fine-tuning. Our code is provided at https://github.com/Luo-Yihong/YOSO.

comment: Revision

♻ ☆ Enhanced Prompt-leveraged Weakly Supervised Cancer Segmentation based on Segment Anything

Joonhyeon Song, Seohwan Yun, Seongho Yoon, Joohyeok Kim, Sangmin Lee

This work proposes a novel approach beyond supervised learning for effective pathological image analysis, addressing the challenge of limited robust labeled data. Pathological diagnosis of diseases like cancer has conventionally relied on the evaluation of morphological features by physicians and pathologists. However, recent advancements in compute-aided diagnosis (CAD) systems are gaining significant attention as diagnostic support tools. Although the advancement of deep learning has improved CAD significantly, segmentation models typically require large pixel-level annotated dataset, and such labeling is expensive. Existing studies not based on supervised approaches still struggle with limited generalization, and no practical approach has emerged yet. To address this issue, we present a weakly supervised semantic segmentation (WSSS) model by combining class activation map and Segment Anything Model (SAM)-based pseudo-labeling. For effective pretraining, we adopt the SAM-a foundation model that is pretrained on large datasets and operates in zero-shot configurations using only coarse prompts. The proposed approach transfer enhanced Attention Dropout Layer's knowledge to SAM, thereby generating pseudo-labels. To demonstrate the superiority of the proposed method, experimental studies are conducted on histopathological breast cancer datasets. The proposed method outperformed other WSSS methods across three datasets, demonstrating its efficiency by achieving this with only 12GB of GPU memory during training. Our code is available at : https://github.com/QI-NemoSong/EPLC-SAM

comment: 10 pages, 7 figures

♻ ☆ Look, Listen, and Answer: Overcoming Biases for Audio-Visual Question Answering NeurIPS 2024

Jie Ma, Min Hu, Pinghui Wang, Wangchun Sun, Lingyun Song, Hongbin Pei, Jun Liu, Youtian Du

Audio-Visual Question Answering (AVQA) is a complex multi-modal reasoning task, demanding intelligent systems to accurately respond to natural language queries based on audio-video input pairs. Nevertheless, prevalent AVQA approaches are prone to overlearning dataset biases, resulting in poor robustness. Furthermore, current datasets may not provide a precise diagnostic for these methods. To tackle these challenges, firstly, we propose a novel dataset, MUSIC-AVQA-R, crafted in two steps: rephrasing questions within the test split of a public dataset (MUSIC-AVQA) and subsequently introducing distribution shifts to split questions. The former leads to a large, diverse test space, while the latter results in a comprehensive robustness evaluation on rare, frequent, and overall questions. Secondly, we propose a robust architecture that utilizes a multifaceted cycle collaborative debiasing strategy to overcome bias learning. Experimental results show that this architecture achieves state-of-the-art performance on MUSIC-AVQA-R, notably obtaining a significant improvement of 9.32%. Extensive ablation experiments are conducted on the two datasets mentioned to analyze the component effectiveness within the debiasing strategy. Additionally, we highlight the limited robustness of existing multi-modal QA methods through the evaluation on our dataset. We also conduct experiments combining various baselines with our proposed strategy on two datasets to verify its plug-and-play capability. Our dataset and code are available at https://github.com/reml-group/MUSIC-AVQA-R.

comment: Accepted by NeurIPS 2024

♻ ☆ NutrifyAI: An AI-Powered System for Real-Time Food Detection, Nutritional Analysis, and Personalized Meal Recommendations

Michelle Han, Junyao Chen, Zhengyuan Zhou

With diet and nutrition apps reaching 1.4 billion users in 2022 [1], it's not surprise that popular health apps, MyFitnessPal, Noom, and Calorie Counter, are surging in popularity. However, one major setback [2] of nearly all nutrition applications is that users must enter food data manually, which is time-consuming and tedious. Thus, there has been an increasing demand for applications that can accurately identify food items, analyze their nutritional content, and offer dietary recommendations in real-time. This paper introduces a comprehensive system that combines advanced computer vision techniques with nutritional analysis, implemented in a versatile mobile and web application. The system is divided into three key concepts: 1) food detection using the YOLOv8 model, 2) nutrient analysis via the Edamam Nutrition Analysis API, and 3) personalized meal recommendations using the Edamam Meal Planning and Recipe Search APIs. Preliminary results showcase the system's effectiveness by providing immediate, accurate dietary insights, with a demonstrated food recognition accuracy of nearly 80%, making it a valuable tool for users to make informed dietary decisions.

comment: 4 pages, 8 figures

♻ ☆ HiRT: Enhancing Robotic Control with Hierarchical Robot Transformers

Jianke Zhang, Yanjiang Guo, Xiaoyu Chen, Yen-Jen Wang, Yucheng Hu, Chengming Shi, Jianyu Chen

Large Vision-Language-Action (VLA) models, leveraging powerful pre trained Vision-Language Models (VLMs) backends, have shown promise in robotic control due to their impressive generalization ability. However, the success comes at a cost. Their reliance on VLM backends with billions of parameters leads to high computational costs and inference latency, limiting the testing scenarios to mainly quasi-static tasks and hindering performance in dynamic tasks requiring rapid interactions. To address these limitations, this paper proposes HiRT, a Hierarchical Robot Transformer framework that enables flexible frequency and performance trade-off. HiRT keeps VLMs running at low frequencies to capture temporarily invariant features while enabling real-time interaction through a high-frequency vision-based policy guided by the slowly updated features. Experiment results in both simulation and real-world settings demonstrate significant improvements over baseline methods. Empirically, in static tasks, we double the control frequency and achieve comparable success rates. Additionally, on novel real-world dynamic ma nipulation tasks which are challenging for previous VLA models, HiRT improves the success rate from 48% to 75%.

♻ ☆ PointSeg: A Training-Free Paradigm for 3D Scene Segmentation via Foundation Models

Qingdong He, Jinlong Peng, Zhengkai Jiang, Xiaobin Hu, Jiangning Zhang, Qiang Nie, Yabiao Wang, Chengjie Wang

Recent success of vision foundation models have shown promising performance for the 2D perception tasks. However, it is difficult to train a 3D foundation network directly due to the limited dataset and it remains under explored whether existing foundation models can be lifted to 3D space seamlessly. In this paper, we present PointSeg, a novel training-free paradigm that leverages off-the-shelf vision foundation models to address 3D scene perception tasks. PointSeg can segment anything in 3D scene by acquiring accurate 3D prompts to align their corresponding pixels across frames. Concretely, we design a two-branch prompts learning structure to construct the 3D point-box prompts pairs, combining with the bidirectional matching strategy for accurate point and proposal prompts generation. Then, we perform the iterative post-refinement adaptively when cooperated with different vision foundation models. Moreover, we design a affinity-aware merging algorithm to improve the final ensemble masks. PointSeg demonstrates impressive segmentation performance across various datasets, all without training. Specifically, our approach significantly surpasses the state-of-the-art specialist training-free model by 14.1$\%$, 12.3$\%$, and 12.6$\%$ mAP on ScanNet, ScanNet++, and KITTI-360 datasets, respectively. On top of that, PointSeg can incorporate with various foundation models and even surpasses the specialist training-based methods by 3.4$\%$-5.4$\%$ mAP across various datasets, serving as an effective generalist model.

♻ ☆ LiteVLoc: Map-Lite Visual Localization for Image Goal Navigation

Jianhao Jiao, Jinhao He, Changkun Liu, Sebastian Aegidius, Xiangcheng Hu, Tristan Braud, Dimitrios Kanoulas

This paper presents LiteVLoc, a hierarchical visual localization framework that uses a lightweight topo-metric map to represent the environment. The method consists of three sequential modules that estimate camera poses in a coarse-to-fine manner. Unlike mainstream approaches relying on detailed 3D representations, LiteVLoc reduces storage overhead by leveraging learning-based feature matching and geometric solvers for metric pose estimation. A novel dataset for the map-free relocalization task is also introduced. Extensive experiments including localization and navigation in both simulated and real-world scenarios have validate the system's performance and demonstrated its precision and efficiency for large-scale deployment. Code and data will be made publicly available.

comment: 9 pages, 4 figures

♻ ☆ Cardiac Copilot: Automatic Probe Guidance for Echocardiography with World Model MICCAI2024

Haojun Jiang, Zhenguo Sun, Ning Jia, Meng Li, Yu Sun, Shaqi Luo, Shiji Song, Gao Huang

Echocardiography is the only technique capable of real-time imaging of the heart and is vital for diagnosing the majority of cardiac diseases. However, there is a severe shortage of experienced cardiac sonographers, due to the heart's complex structure and significant operational challenges. To mitigate this situation, we present a Cardiac Copilot system capable of providing real-time probe movement guidance to assist less experienced sonographers in conducting freehand echocardiography. This system can enable non-experts, especially in primary departments and medically underserved areas, to perform cardiac ultrasound examinations, potentially improving global healthcare delivery. The core innovation lies in proposing a data-driven world model, named Cardiac Dreamer, for representing cardiac spatial structures. This world model can provide structure features of any cardiac planes around the current probe position in the latent space, serving as an precise navigation map for autonomous plane localization. We train our model with real-world ultrasound data and corresponding probe motion from 110 routine clinical scans with 151K sample pairs by three certified sonographers. Evaluations on three standard planes with 37K sample pairs demonstrate that the world model can reduce navigation errors by up to 33\% and exhibit more stable performance.

comment: Accepted by MICCAI2024

♻ ☆ A Rainbow in Deep Network Black Boxes

Florentin Guth, Brice Ménard, Gaspar Rochette, Stéphane Mallat

A central question in deep learning is to understand the functions learned by deep networks. What is their approximation class? Do the learned weights and representations depend on initialization? Previous empirical work has evidenced that kernels defined by network activations are similar across initializations. For shallow networks, this has been theoretically studied with random feature models, but an extension to deep networks has remained elusive. Here, we provide a deep extension of such random feature models, which we call the rainbow model. We prove that rainbow networks define deterministic (hierarchical) kernels in the infinite-width limit. The resulting functions thus belong to a data-dependent RKHS which does not depend on the weight randomness. We also verify numerically our modeling assumptions on deep CNNs trained on image classification tasks, and show that the trained networks approximately satisfy the rainbow hypothesis. In particular, rainbow networks sampled from the corresponding random feature model achieve similar performance as the trained networks. Our results highlight the central role played by the covariances of network weights at each layer, which are observed to be low-rank as a result of feature learning.

comment: 59 pages, 10 figures. To appear at JMLR

♻ ☆ FSL-Rectifier: Rectify Outliers in Few-Shot Learning via Test-Time Augmentation

Yunwei Bai, Ying Kiat Tan, Shiming Chen, Yao Shu, Tsuhan Chen

Few-shot-learning (FSL) commonly requires a model to identify images (queries) that belong to classes unseen during training, based on a few labeled samples of the new classes (support set) as reference. So far, plenty of algorithms involve training data augmentation to improve the generalization capability of FSL models, but outlier queries or support images during inference can still pose great generalization challenges. In this work, to reduce the bias caused by the outlier samples, we generate additional test-class samples by combining original samples with suitable train-class samples via a generative image combiner. Then, we obtain averaged features via an augmentor, which leads to more typical representations through the averaging. We experimentally and theoretically demonstrate the effectiveness of our method, e.g., obtaining a test accuracy improvement proportion of around 10% (e.g., from 46.86% to 53.28%) for trained FSL models. Importantly, given pretrained image combiner, our method is training-free for off-the-shelf FSL models, whose performance can be improved without extra datasets nor further training of the models themselves.

♻ ☆ GMAI-MMBench: A Comprehensive Multimodal Evaluation Benchmark Towards General Medical AI

Pengcheng Chen, Jin Ye, Guoan Wang, Yanjun Li, Zhongying Deng, Wei Li, Tianbin Li, Haodong Duan, Ziyan Huang, Yanzhou Su, Benyou Wang, Shaoting Zhang, Bin Fu, Jianfei Cai, Bohan Zhuang, Eric J Seibel, Junjun He, Yu Qiao

Large Vision-Language Models (LVLMs) are capable of handling diverse data types such as imaging, text, and physiological signals, and can be applied in various fields. In the medical field, LVLMs have a high potential to offer substantial assistance for diagnosis and treatment. Before that, it is crucial to develop benchmarks to evaluate LVLMs' effectiveness in various medical applications. Current benchmarks are often built upon specific academic literature, mainly focusing on a single domain, and lacking varying perceptual granularities. Thus, they face specific challenges, including limited clinical relevance, incomplete evaluations, and insufficient guidance for interactive LVLMs. To address these limitations, we developed the GMAI-MMBench, the most comprehensive general medical AI benchmark with well-categorized data structure and multi-perceptual granularity to date. It is constructed from 284 datasets across 38 medical image modalities, 18 clinical-related tasks, 18 departments, and 4 perceptual granularities in a Visual Question Answering (VQA) format. Additionally, we implemented a lexical tree structure that allows users to customize evaluation tasks, accommodating various assessment needs and substantially supporting medical AI research and applications. We evaluated 50 LVLMs, and the results show that even the advanced GPT-4o only achieves an accuracy of 53.96%, indicating significant room for improvement. Moreover, we identified five key insufficiencies in current cutting-edge LVLMs that need to be addressed to advance the development of better medical applications. We believe that GMAI-MMBench will stimulate the community to build the next generation of LVLMs toward GMAI.

comment: GitHub: https://github.com/uni-medical/GMAI-MMBench Hugging face: https://huggingface.co/datasets/OpenGVLab/GMAI-MMBench

♻ ☆ Open-World Continual Learning: Unifying Novelty Detection and Continual Learning

Gyuhak Kim, Changnan Xiao, Tatsuya Konishi, Zixuan Ke, Bing Liu

As AI agents are increasingly used in the real open world with unknowns or novelties, they need the ability to (1) recognize objects that (a) they have learned before and (b) detect items that they have never seen or learned, and (2) learn the new items incrementally to become more and more knowledgeable and powerful. (1) is called novelty detection or out-of-distribution (OOD) detection and (2) is called class incremental learning (CIL), which is a setting of continual learning (CL). In existing research, OOD detection and CIL are regarded as two completely different problems. This paper first provides a theoretical proof that good OOD detection for each task within the set of learned tasks (called closed-world OOD detection) is necessary for successful CIL. We show this by decomposing CIL into two sub-problems: within-task prediction (WP) and task-id prediction (TP), and proving that TP is correlated with closed-world OOD detection. The key theoretical result is that regardless of whether WP and OOD detection (or TP) are defined explicitly or implicitly by a CIL algorithm, good WP and good closed-world OOD detection are necessary and sufficient conditions for good CIL, which unifies novelty or OOD detection and continual learning (CIL, in particular). We call this traditional CIL the closed-world CIL as it does not detect future OOD data in the open world. The paper then proves that the theory can be generalized or extended to open-world CIL, which is the proposed open-world continual learning, that can perform CIL in the open world and detect future or open-world OOD data. Based on the theoretical results, new CIL methods are also designed, which outperform strong baselines in CIL accuracy and in continual OOD detection by a large margin.

comment: To appear in Artificial Intelligence Journal. arXiv admin note: substantial text overlap with arXiv:2211.02633

♻ ☆ PIR: Remote Sensing Image-Text Retrieval with Prior Instruction Representation Learning

Jiancheng Pan, Muyuan Ma, Qing Ma, Cong Bai, Shengyong Chen

Remote sensing image-text retrieval constitutes a foundational aspect of remote sensing interpretation tasks, facilitating the alignment of vision and language representations. This paper introduces a prior instruction representation (PIR) learning paradigm that draws on prior knowledge to instruct adaptive learning of vision and text representations. Based on PIR, a domain-adapted remote sensing image-text retrieval framework PIR-ITR is designed to address semantic noise issues in vision-language understanding tasks. However, with massive additional data for pre-training the vision-language foundation model, remote sensing image-text retrieval is further developed into an open-domain retrieval task. Continuing with the above, we propose PIR-CLIP, a domain-specific CLIP-based framework for remote sensing image-text retrieval, to address semantic noise in remote sensing vision-language representations and further improve open-domain retrieval performance. In vision representation, we utilize the prior-guided knowledge of the remote sensing scene recognition by building a belief matrix to select key features for reducing the impact of semantic noise. In text representation, we use the previous time step to cyclically activate the current time step to enhance text representation capability. A cluster-wise Affiliation Loss (AL) is proposed to constrain the inter-classes and to reduce the semantic confusion zones in the common subspace. Comprehensive experiments demonstrate that PIR could enhance vision and text representations and outperform the state-of-the-art methods of closed-domain and open-domain retrieval on two benchmark datasets, RSICD and RSITMD.

comment: 13 pages, 8 figures

♻ ☆ MAL: Motion-Aware Loss with Temporal and Distillation Hints for Self-Supervised Depth Estimation ICRA 2024

Yue-Jiang Dong, Fang-Lue Zhang, Song-Hai Zhang

Depth perception is crucial for a wide range of robotic applications. Multi-frame self-supervised depth estimation methods have gained research interest due to their ability to leverage large-scale, unlabeled real-world data. However, the self-supervised methods often rely on the assumption of a static scene and their performance tends to degrade in dynamic environments. To address this issue, we present Motion-Aware Loss, which leverages the temporal relation among consecutive input frames and a novel distillation scheme between the teacher and student networks in the multi-frame self-supervised depth estimation methods. Specifically, we associate the spatial locations of moving objects with the temporal order of input frames to eliminate errors induced by object motion. Meanwhile, we enhance the original distillation scheme in multi-frame methods to better exploit the knowledge from a teacher network. MAL is a novel, plug-and-play module designed for seamless integration into multi-frame self-supervised monocular depth estimation methods. Adding MAL into previous state-of-the-art methods leads to a reduction in depth estimation errors by up to 4.2% and 10.8% on KITTI and CityScapes benchmarks, respectively.

comment: Accepted by ICRA 2024; Project homepage: https://yuejiangdong.github.io/MotionAwareLoss/

♻ ☆ End-to-End Rate-Distortion Optimized 3D Gaussian Representation ECCV 2024

Henan Wang, Hanxin Zhu, Tianyu He, Runsen Feng, Jiajun Deng, Jiang Bian, Zhibo Chen

3D Gaussian Splatting (3DGS) has become an emerging technique with remarkable potential in 3D representation and image rendering. However, the substantial storage overhead of 3DGS significantly impedes its practical applications. In this work, we formulate the compact 3D Gaussian learning as an end-to-end Rate-Distortion Optimization (RDO) problem and propose RDO-Gaussian that can achieve flexible and continuous rate control. RDO-Gaussian addresses two main issues that exist in current schemes: 1) Different from prior endeavors that minimize the rate under the fixed distortion, we introduce dynamic pruning and entropy-constrained vector quantization (ECVQ) that optimize the rate and distortion at the same time. 2) Previous works treat the colors of each Gaussian equally, while we model the colors of different regions and materials with learnable numbers of parameters. We verify our method on both real and synthetic scenes, showcasing that RDO-Gaussian greatly reduces the size of 3D Gaussian over 40x, and surpasses existing methods in rate-distortion performance.

comment: ECCV 2024

♻ ☆ CinePile: A Long Video Question Answering Dataset and Benchmark

Ruchit Rawal, Khalid Saifullah, Miquel Farré, Ronen Basri, David Jacobs, Gowthami Somepalli, Tom Goldstein

Current datasets for long-form video understanding often fall short of providing genuine long-form comprehension challenges, as many tasks derived from these datasets can be successfully tackled by analyzing just one or a few random frames from a video. To address this issue, we present a novel dataset and benchmark, CinePile, specifically designed for authentic long-form video understanding. This paper details our innovative approach for creating a question-answer dataset, utilizing advanced LLMs with human-in-the-loop and building upon human-generated raw data. Our comprehensive dataset comprises 305,000 multiple-choice questions (MCQs), covering various visual and multimodal aspects, including temporal comprehension, understanding human-object interactions, and reasoning about events or actions within a scene. Additionally, we fine-tuned open-source Video-LLMs on the training split and evaluated both open-source and proprietary video-centric LLMs on the test split of our dataset. The findings indicate that although current models underperform compared to humans, fine-tuning these models can lead to significant improvements in their performance.

comment: Project page with all the artifacts - https://ruchitrawal.github.io/cinepile/. Updated version with adversarial refinement pipeline and more model evaluations

♻ ★ Show-o: One Single Transformer to Unify Multimodal Understanding and Generation

Jinheng Xie, Weijia Mao, Zechen Bai, David Junhao Zhang, Weihao Wang, Kevin Qinghong Lin, Yuchao Gu, Zhijie Chen, Zhenheng Yang, Mike Zheng Shou

We present a unified transformer, i.e., Show-o, that unifies multimodal understanding and generation. Unlike fully autoregressive models, Show-o unifies autoregressive and (discrete) diffusion modeling to adaptively handle inputs and outputs of various and mixed modalities. The unified model flexibly supports a wide range of vision-language tasks including visual question-answering, text-to-image generation, text-guided inpainting/extrapolation, and mixed-modality generation. Across various benchmarks, it demonstrates comparable or superior performance to existing individual models with an equivalent or larger number of parameters tailored for understanding or generation. This significantly highlights its potential as a next-generation foundation model. Code and models are released at https://github.com/showlab/Show-o.

comment: Technical Report

☆ Reflection-Bench: probing AI intelligence with reflection

Lingyu Li, Yixu Wang, Haiquan Zhao, Shuqi Kong, Yan Teng, Chunbo Li, Yingchun Wang

The ability to adapt beliefs or behaviors in response to unexpected outcomes, reflection, is fundamental to intelligent systems' interaction with the world. From a cognitive science perspective, this serves as a core principle of intelligence applicable to both human and AI systems. To address the debate on the intelligence of large language models (LLMs), we propose Reflection-Bench, a comprehensive benchmark comprising 7 tasks spanning core cognitive functions crucial for reflection, including perception, memory, belief updating, decision-making, prediction, counterfactual thinking, and meta-reflection. We evaluate the performances of 13 prominent LLMs such as OpenAI o1, GPT-4, Claude 3.5 Sonnet, etc. The results indicate that current LLMs still lack satisfactory reflection ability. We discuss the underlying causes of these results and suggest potential avenues for future research. In conclusion, Reflection-Bench offers both evaluation tools and inspiration for developing AI capable of reliably interacting with the environment. Our data and code are available at https://github.com/YabYum/ReflectionBench.

comment: 11 pages, 7 figures, 2 tables

☆ xGen-MM-Vid (BLIP-3-Video): You Only Need 32 Tokens to Represent a Video Even in VLMs

Michael S. Ryoo, Honglu Zhou, Shrikant Kendre, Can Qin, Le Xue, Manli Shu, Silvio Savarese, Ran Xu, Caiming Xiong, Juan Carlos Niebles

We present xGen-MM-Vid (BLIP-3-Video): a multimodal language model for videos, particularly designed to efficiently capture temporal information over multiple frames. BLIP-3-Video takes advantage of the 'temporal encoder' in addition to the conventional visual tokenizer, which maps a sequence of tokens over multiple frames into a compact set of visual tokens. This enables BLIP3-Video to use much fewer visual tokens than its competing models (e.g., 32 vs. 4608 tokens). We explore different types of temporal encoders, including learnable spatio-temporal pooling as well as sequential models like Token Turing Machines. We experimentally confirm that BLIP-3-Video obtains video question-answering accuracies comparable to much larger state-of-the-art models (e.g., 34B), while being much smaller (i.e., 4B) and more efficient by using fewer visual tokens. The project website is at https://www.salesforceairesearch.com/opensource/xGen-MM-Vid/index.html

☆ 3DGS-Enhancer: Enhancing Unbounded 3D Gaussian Splatting with View-consistent 2D Diffusion Priors NeurIPS 2024

Xi Liu, Chaoyi Zhou, Siyu Huang

Novel-view synthesis aims to generate novel views of a scene from multiple input images or videos, and recent advancements like 3D Gaussian splatting (3DGS) have achieved notable success in producing photorealistic renderings with efficient pipelines. However, generating high-quality novel views under challenging settings, such as sparse input views, remains difficult due to insufficient information in under-sampled areas, often resulting in noticeable artifacts. This paper presents 3DGS-Enhancer, a novel pipeline for enhancing the representation quality of 3DGS representations. We leverage 2D video diffusion priors to address the challenging 3D view consistency problem, reformulating it as achieving temporal consistency within a video generation process. 3DGS-Enhancer restores view-consistent latent features of rendered novel views and integrates them with the input views through a spatial-temporal decoder. The enhanced views are then used to fine-tune the initial 3DGS model, significantly improving its rendering performance. Extensive experiments on large-scale datasets of unbounded scenes demonstrate that 3DGS-Enhancer yields superior reconstruction performance and high-fidelity rendering results compared to state-of-the-art methods. The project webpage is https://xiliu8006.github.io/3DGS-Enhancer-project .

comment: Accepted by NeurIPS 2024 Spotlight

☆ CompassJudger-1: All-in-one Judge Model Helps Model Evaluation and Evolution

Maosong Cao, Alexander Lam, Haodong Duan, Hongwei Liu, Songyang Zhang, Kai Chen

Efficient and accurate evaluation is crucial for the continuous improvement of large language models (LLMs). Among various assessment methods, subjective evaluation has garnered significant attention due to its superior alignment with real-world usage scenarios and human preferences. However, human-based evaluations are costly and lack reproducibility, making precise automated evaluators (judgers) vital in this process. In this report, we introduce \textbf{CompassJudger-1}, the first open-source \textbf{all-in-one} judge LLM. CompassJudger-1 is a general-purpose LLM that demonstrates remarkable versatility. It is capable of: 1. Performing unitary scoring and two-model comparisons as a reward model; 2. Conducting evaluations according to specified formats; 3. Generating critiques; 4. Executing diverse tasks like a general LLM. To assess the evaluation capabilities of different judge models under a unified setting, we have also established \textbf{JudgerBench}, a new benchmark that encompasses various subjective evaluation tasks and covers a wide range of topics. CompassJudger-1 offers a comprehensive solution for various evaluation tasks while maintaining the flexibility to adapt to diverse requirements. Both CompassJudger and JudgerBench are released and available to the research community athttps://github.com/open-compass/CompassJudger. We believe that by open-sourcing these tools, we can foster collaboration and accelerate progress in LLM evaluation methodologies.

comment: Technical Report, Code and Models: https://github.com/open-compass/CompassJudger

☆ MoRE: Multi-Modal Contrastive Pre-training with Transformers on X-Rays, ECGs, and Diagnostic Report

Samrajya Thapa, Koushik Howlader, Subhankar Bhattacharjee, Wei le

In this paper, we introduce a novel Multi-Modal Contrastive Pre-training Framework that synergistically combines X-rays, electrocardiograms (ECGs), and radiology/cardiology reports. Our approach leverages transformers to encode these diverse modalities into a unified representation space, aiming to enhance diagnostic accuracy and facilitate comprehensive patient assessments. We utilize LoRA-Peft to significantly reduce trainable parameters in the LLM and incorporate recent linear attention dropping strategy in the Vision Transformer(ViT) for smoother attention. Furthermore, we provide novel multimodal attention explanations and retrieval for our model. To the best of our knowledge, we are the first to propose an integrated model that combines X-ray, ECG, and Radiology/Cardiology Report with this approach. By utilizing contrastive loss, MoRE effectively aligns modality-specific features into a coherent embedding, which supports various downstream tasks such as zero-shot classification and multimodal retrieval. Employing our proposed methodology, we achieve state-of-the-art (SOTA) on the Mimic-IV, CheXpert, Edema Severity, and PtbXl downstream datasets, surpassing existing multimodal approaches. Our proposed framework shows significant improvements in capturing intricate inter-modal relationships and its robustness in medical diagnosis that establishes a framework for future research in multimodal learning in the healthcare sector.

comment: 10 pages, 5 figures, 9 tables. Supplementary detail in Appendix. Code made available in Github for reproducibility

☆ Sketch2Code: Evaluating Vision-Language Models for Interactive Web Design Prototyping

Ryan Li, Yanzhe Zhang, Diyi Yang

Sketches are a natural and accessible medium for UI designers to conceptualize early-stage ideas. However, existing research on UI/UX automation often requires high-fidelity inputs like Figma designs or detailed screenshots, limiting accessibility and impeding efficient design iteration. To bridge this gap, we introduce Sketch2Code, a benchmark that evaluates state-of-the-art Vision Language Models (VLMs) on automating the conversion of rudimentary sketches into webpage prototypes. Beyond end-to-end benchmarking, Sketch2Code supports interactive agent evaluation that mimics real-world design workflows, where a VLM-based agent iteratively refines its generations by communicating with a simulated user, either passively receiving feedback instructions or proactively asking clarification questions. We comprehensively analyze ten commercial and open-source models, showing that Sketch2Code is challenging for existing VLMs; even the most capable models struggle to accurately interpret sketches and formulate effective questions that lead to steady improvement. Nevertheless, a user study with UI/UX experts reveals a significant preference for proactive question-asking over passive feedback reception, highlighting the need to develop more effective paradigms for multi-turn conversational agents.

comment: preprint, 9 pages

☆ Pre-training Distillation for Large Language Models: A Design Space Exploration

Hao Peng, Xin Lv, Yushi Bai, Zijun Yao, Jiajie Zhang, Lei Hou, Juanzi Li

Knowledge distillation (KD) aims to transfer knowledge from a large teacher model to a smaller student model. Previous work applying KD in the field of large language models (LLMs) typically focused on the post-training phase, where the student LLM learns directly from instructions and corresponding responses generated by the teacher model. In this paper, we extend KD to the pre-training phase of LLMs, named pre-training distillation (PD). We first conduct a preliminary experiment using GLM-4-9B as the teacher LLM to distill a 1.9B parameter student LLM, validating the effectiveness of PD. Considering the key impact factors of distillation, we systematically explore the design space of pre-training distillation across four aspects: logits processing, loss selection, scaling law, and offline or online logits. We conduct extensive experiments to explore the design space of pre-training distillation and find better configurations and interesting conclusions, such as larger student LLMs generally benefiting more from pre-training distillation, while a larger teacher LLM does not necessarily guarantee better results. We hope our exploration of the design space will inform future practices in pre-training distillation.

☆ Comprehensive benchmarking of large language models for RNA secondary structure prediction

L. I. Zablocki, L. A. Bugnon, M. Gerard, L. Di Persia, G. Stegmayer, D. H. Milone

Inspired by the success of large language models (LLM) for DNA and proteins, several LLM for RNA have been developed recently. RNA-LLM uses large datasets of RNA sequences to learn, in a self-supervised way, how to represent each RNA base with a semantically rich numerical vector. This is done under the hypothesis that obtaining high-quality RNA representations can enhance data-costly downstream tasks. Among them, predicting the secondary structure is a fundamental task for uncovering RNA functional mechanisms. In this work we present a comprehensive experimental analysis of several pre-trained RNA-LLM, comparing them for the RNA secondary structure prediction task in an unified deep learning framework. The RNA-LLM were assessed with increasing generalization difficulty on benchmark datasets. Results showed that two LLM clearly outperform the other models, and revealed significant challenges for generalization in low-homology scenarios.

☆ Compute-Constrained Data Selection

Junjie Oscar Yin, Alexander M. Rush

Data selection can reduce the amount of training data needed to finetune LLMs; however, the efficacy of data selection scales directly with its compute. Motivated by the practical challenge of compute-constrained finetuning, we consider the setting in which both the cost of selecting data and training are budgeted for. We first formalize the problem of data selection with a cost-aware utility function, and model the data selection problem as trading off initial-selection cost for training gain. We run a comprehensive sweep of experiments across multiple tasks, varying compute budget by scaling finetuning tokens, model sizes, and data selection compute. These experiments show the validity of this model in real-world experiments. Interestingly we find that many powerful data selection methods are almost never compute-optimal, and that cheaper data selection alternatives dominate both from a theoretical and empirical perspective.

☆ Improve Vision Language Model Chain-of-thought Reasoning

Ruohong Zhang, Bowen Zhang, Yanghao Li, Haotian Zhang, Zhiqing Sun, Zhe Gan, Yinfei Yang, Ruoming Pang, Yiming Yang

Chain-of-thought (CoT) reasoning in vision language models (VLMs) is crucial for improving interpretability and trustworthiness. However, current training recipes lack robust CoT reasoning data, relying on datasets dominated by short annotations with minimal rationales. In this work, we show that training VLM on short answers does not generalize well to reasoning tasks that require more detailed responses. To address this, we propose a two-fold approach. First, we distill rationales from GPT-4o model to enrich the training data and fine-tune VLMs, boosting their CoT performance. Second, we apply reinforcement learning to further calibrate reasoning quality. Specifically, we construct positive (correct) and negative (incorrect) pairs of model-generated reasoning chains, by comparing their predictions with annotated short answers. Using this pairwise data, we apply the Direct Preference Optimization algorithm to refine the model's reasoning abilities. Our experiments demonstrate significant improvements in CoT reasoning on benchmark datasets and better generalization to direct answer prediction as well. This work emphasizes the importance of incorporating detailed rationales in training and leveraging reinforcement learning to strengthen the reasoning capabilities of VLMs.

comment: 10 pages + appendix

☆ Information for Conversation Generation: Proposals Utilising Knowledge Graphs ISWC 2024

Alex Clay, Ernesto Jiménez-Ruiz

LLMs are frequently used tools for conversational generation. Without additional information LLMs can generate lower quality responses due to lacking relevant content and hallucinations, as well as the perception of poor emotional capability, and an inability to maintain a consistent character. Knowledge graphs are commonly used forms of external knowledge and may provide solutions to these challenges. This paper introduces three proposals, utilizing knowledge graphs to enhance LLM generation. Firstly, dynamic knowledge graph embeddings and recommendation could allow for the integration of new information and the selection of relevant knowledge for response generation. Secondly, storing entities with emotional values as additional features may provide knowledge that is better emotionally aligned with the user input. Thirdly, integrating character information through narrative bubbles would maintain character consistency, as well as introducing a structure that would readily incorporate new information.

comment: 7 pages with citations, 1 figure, accepted to the ISWC 2024 Special Session

☆ Learning How to Vote With Principles: Axiomatic Insights Into the Collective Decisions of Neural Networks

Levin Hornischer, Zoi Terzopoulou

Can neural networks be applied in voting theory, while satisfying the need for transparency in collective decisions? We propose axiomatic deep voting: a framework to build and evaluate neural networks that aggregate preferences, using the well-established axiomatic method of voting theory. Our findings are: (1) Neural networks, despite being highly accurate, often fail to align with the core axioms of voting rules, revealing a disconnect between mimicking outcomes and reasoning. (2) Training with axiom-specific data does not enhance alignment with those axioms. (3) By solely optimizing axiom satisfaction, neural networks can synthesize new voting rules that often surpass and substantially differ from existing ones. This offers insights for both fields: For AI, important concepts like bias and value-alignment are studied in a mathematically rigorous way; for voting theory, new areas of the space of voting rules are explored.

comment: 15 pages, 8 figures, 7 tables

☆ GenAI Assisting Medical Training

Stefan Fritsch, Matthias Tschoepe, Vitor Fortes Rey, Lars Krupp, Agnes Gruenerbl, Eloise Monger, Sarah Travenna

Medical procedures such as venipuncture and cannulation are essential for nurses and require precise skills. Learning this skill, in turn, is a challenge for educators due to the number of teachers per class and the complexity of the task. The study aims to help students with skill acquisition and alleviate the educator's workload by integrating generative AI methods to provide real-time feedback on medical procedures such as venipuncture and cannulation.

comment: 2 pages, 2 figures

☆ Warped Diffusion: Solving Video Inverse Problems with Image Diffusion Models NeurIPS 2024

Giannis Daras, Weili Nie, Karsten Kreis, Alex Dimakis, Morteza Mardani, Nikola Borislavov Kovachki, Arash Vahdat

Using image models naively for solving inverse video problems often suffers from flickering, texture-sticking, and temporal inconsistency in generated videos. To tackle these problems, in this paper, we view frames as continuous functions in the 2D space, and videos as a sequence of continuous warping transformations between different frames. This perspective allows us to train function space diffusion models only on images and utilize them to solve temporally correlated inverse problems. The function space diffusion models need to be equivariant with respect to the underlying spatial transformations. To ensure temporal consistency, we introduce a simple post-hoc test-time guidance towards (self)-equivariant solutions. Our method allows us to deploy state-of-the-art latent diffusion models such as Stable Diffusion XL to solve video inverse problems. We demonstrate the effectiveness of our method for video inpainting and $8\times$ video super-resolution, outperforming existing techniques based on noise transformations. We provide generated video results: https://giannisdaras.github.io/warped\_diffusion.github.io/.

comment: Accepted in NeurIPS 2024

☆ Small Contributions, Small Networks: Efficient Neural Network Pruning Based on Relative Importance

Mostafa Hussien, Mahmoud Afifi, Kim Khoa Nguyen, Mohamed Cheriet

Recent advancements have scaled neural networks to unprecedented sizes, achieving remarkable performance across a wide range of tasks. However, deploying these large-scale models on resource-constrained devices poses significant challenges due to substantial storage and computational requirements. Neural network pruning has emerged as an effective technique to mitigate these limitations by reducing model size and complexity. In this paper, we introduce an intuitive and interpretable pruning method based on activation statistics, rooted in information theory and statistical analysis. Our approach leverages the statistical properties of neuron activations to identify and remove weights with minimal contributions to neuron outputs. Specifically, we build a distribution of weight contributions across the dataset and utilize its parameters to guide the pruning process. Furthermore, we propose a Pruning-aware Training strategy that incorporates an additional regularization term to enhance the effectiveness of our pruning method. Extensive experiments on multiple datasets and network architectures demonstrate that our method consistently outperforms several baseline and state-of-the-art pruning techniques.

☆ PODTILE: Facilitating Podcast Episode Browsing with Auto-generated Chapters CIKM

Azin Ghazimatin, Ekaterina Garmash, Gustavo Penha, Kristen Sheets, Martin Achenbach, Oguz Semerci, Remi Galvez, Marcus Tannenberg, Sahitya Mantravadi, Divya Narayanan, Ofeliya Kalaydzhyan, Douglas Cole, Ben Carterette, Ann Clifton, Paul N. Bennett, Claudia Hauff, Mounia Lalmas

Listeners of long-form talk-audio content, such as podcast episodes, often find it challenging to understand the overall structure and locate relevant sections. A practical solution is to divide episodes into chapters--semantically coherent segments labeled with titles and timestamps. Since most episodes on our platform at Spotify currently lack creator-provided chapters, automating the creation of chapters is essential. Scaling the chapterization of podcast episodes presents unique challenges. First, episodes tend to be less structured than written texts, featuring spontaneous discussions with nuanced transitions. Second, the transcripts are usually lengthy, averaging about 16,000 tokens, which necessitates efficient processing that can preserve context. To address these challenges, we introduce PODTILE, a fine-tuned encoder-decoder transformer to segment conversational data. The model simultaneously generates chapter transitions and titles for the input transcript. To preserve context, each input text is augmented with global context, including the episode's title, description, and previous chapter titles. In our intrinsic evaluation, PODTILE achieved an 11% improvement in ROUGE score over the strongest baseline. Additionally, we provide insights into the practical benefits of auto-generated chapters for listeners navigating episode content. Our findings indicate that auto-generated chapters serve as a useful tool for engaging with less popular podcasts. Finally, we present empirical evidence that using chapter titles can enhance effectiveness of sparse retrieval in search tasks.

comment: 9 pages, 4 figures, CIKM industry track 2024

☆ Modeling dynamic neural activity by combining naturalistic video stimuli and stimulus-independent latent factors

Finn Schmidt, Suhas Shrinivasan, Polina Turishcheva, Fabian H. Sinz

Understanding how the brain processes dynamic natural stimuli remains a fundamental challenge in neuroscience. Current dynamic neural encoding models either take stimuli as input but ignore shared variability in neural responses, or they model this variability by deriving latent embeddings from neural responses or behavior while ignoring the visual input. To address this gap, we propose a probabilistic model that incorporates video inputs along with stimulus-independent latent factors to capture variability in neuronal responses, predicting a joint distribution for the entire population. After training and testing our model on mouse V1 neuronal responses, we found that it outperforms video-only models in terms of log-likelihood and achieves further improvements when conditioned on responses from other neurons. Furthermore, we find that the learned latent factors strongly correlate with mouse behavior, although the model was trained without behavior data.

☆ Beyond 2:4: exploring V:N:M sparsity for efficient transformer inference on GPUs

Kang Zhao, Tao Yuan, Han Bao, Zhenfeng Su, Chang Gao, Zhaofeng Sun, Zichen Liang, Liping Jing, Jianfei Chen

To date, 2:4 sparsity has stood as the only sparse pattern that can be accelerated using sparse tensor cores on GPUs. In practice, 2:4 sparsity often possesses low actual speedups ($\leq 1.3$) and requires fixed sparse ratios, meaning that other ratios, such as 4:8, 8:16, or those exceeding 50% sparsity, do not incur any speedups on GPUs. Recent studies suggest that V:N:M sparsity is promising in addressing these limitations of 2:4 sparsity. However, regarding accuracy, the effects of V:N:M sparsity on broader Transformer models, such as vision Transformers and large language models (LLMs), are largely unexamined. Moreover, Some specific issues related to V:N:M sparsity, such as how to select appropriate V and M values, remain unresolved. In this study, we thoroughly investigate the application of V:N:M sparsity in vision models and LLMs across multiple tasks, from pertaining to downstream tasks. We propose three key approaches to enhance the applicability and accuracy of V:N:M-sparse Transformers, including heuristic V and M selection, V:N:M-specific channel permutation, and three-staged LoRA training techniques. Experimental results show that, with our methods, the DeiT-small achieves lossless accuracy at 64:2:5 sparsity, while the DeiT-base maintains accuracy even at 64:2:8 sparsity. In addition, the fine-tuned LLama2-7B at 64:2:5 sparsity performs comparably or better than training-free 2:4 sparse alternatives on downstream tasks. More importantly, V:N:M-sparse Transformers offer a wider range of speedup-accuracy trade-offs compared to 2:4 sparsity. Overall, our exploration largely facilitates the V:N:M sparsity to act as a truly effective acceleration solution for Transformers in cost-sensitive inference scenarios.

☆ A Data-driven Crowd Simulation Framework Integrating Physics-informed Machine Learning with Navigation Potential Fields

Runkang Guo, Bin Chen, Qi Zhang, Yong Zhao, Xiao Wang, Zhengqiu Zhu

Traditional rule-based physical models are limited by their reliance on singular physical formulas and parameters, making it difficult to effectively tackle the intricate tasks associated with crowd simulation. Recent research has introduced deep learning methods to tackle these issues, but most current approaches focus primarily on generating pedestrian trajectories, often lacking interpretability and failing to provide real-time dynamic simulations.To address the aforementioned issues, we propose a novel data-driven crowd simulation framework that integrates Physics-informed Machine Learning (PIML) with navigation potential fields. Our approach leverages the strengths of both physical models and PIML. Specifically, we design an innovative Physics-informed Spatio-temporal Graph Convolutional Network (PI-STGCN) as a data-driven module to predict pedestrian movement trends based on crowd spatio-temporal data. Additionally, we construct a physical model of navigation potential fields based on flow field theory to guide pedestrian movements, thereby reinforcing physical constraints during the simulation. In our framework, navigation potential fields are dynamically computed and updated based on the movement trends predicted by the PI-STGCN, while the updated crowd dynamics, guided by these fields, subsequently feed back into the PI-STGCN. Comparative experiments on two publicly available large-scale real-world datasets across five scenes demonstrate that our proposed framework outperforms existing rule-based methods in accuracy and fidelity. The similarity between simulated and actual pedestrian trajectories increases by 10.8%, while the average error is reduced by 4%. Moreover, our framework exhibits greater adaptability and better interpretability compared to methods that rely solely on deep learning for trajectory generation.

☆ SMART: Self-learning Meta-strategy Agent for Reasoning Tasks

Rongxing Liu, Kumar Shridhar, Manish Prajapat, Patrick Xia, Mrinmaya Sachan

Tasks requiring deductive reasoning, especially those involving multiple steps, often demand adaptive strategies such as intermediate generation of rationales or programs, as no single approach is universally optimal. While Language Models (LMs) can enhance their outputs through iterative self-refinement and strategy adjustments, they frequently fail to apply the most effective strategy in their first attempt. This inefficiency raises the question: Can LMs learn to select the optimal strategy in the first attempt, without a need for refinement? To address this challenge, we introduce SMART (Self-learning Meta-strategy Agent for Reasoning Tasks), a novel framework that enables LMs to autonomously learn and select the most effective strategies for various reasoning tasks. We model the strategy selection process as a Markov Decision Process and leverage reinforcement learning-driven continuous self-improvement to allow the model to find the suitable strategy to solve a given task. Unlike traditional self-refinement methods that rely on multiple inference passes or external feedback, SMART allows an LM to internalize the outcomes of its own reasoning processes and adjust its strategy accordingly, aiming for correct solutions on the first attempt. Our experiments across various reasoning datasets and with different model architectures demonstrate that SMART significantly enhances the ability of models to choose optimal strategies without external guidance (+15 points on the GSM8K dataset). By achieving higher accuracy with a single inference pass, SMART not only improves performance but also reduces computational costs for refinement-based strategies, paving the way for more efficient and intelligent reasoning in LMs.

☆ SeaDAG: Semi-autoregressive Diffusion for Conditional Directed Acyclic Graph Generation

Xinyi Zhou, Xing Li, Yingzhao Lian, Yiwen Wang, Lei Chen, Mingxuan Yuan, Jianye Hao, Guangyong Chen, Pheng Ann Heng

We introduce SeaDAG, a semi-autoregressive diffusion model for conditional generation of Directed Acyclic Graphs (DAGs). Considering their inherent layer-wise structure, we simulate layer-wise autoregressive generation by designing different denoising speed for different layers. Unlike conventional autoregressive generation that lacks a global graph structure view, our method maintains a complete graph structure at each diffusion step, enabling operations such as property control that require the full graph structure. Leveraging this capability, we evaluate the DAG properties during training by employing a graph property decoder. We explicitly train the model to learn graph conditioning with a condition loss, which enhances the diffusion model's capacity to generate graphs that are both realistic and aligned with specified properties. We evaluate our method on two representative conditional DAG generation tasks: (1) circuit generation from truth tables, where precise DAG structures are crucial for realizing circuit functionality, and (2) molecule generation based on quantum properties. Our approach demonstrates promising results, generating high-quality and realistic DAGs that closely align with given conditions.

☆ Multimodal Flare Forecasting with Deep Learning

Grégoire Francisco, Sabrina Guastavino, Teresa Barata, João Fernandes, Dario Del Moro

Solar flare forecasting mainly relies on photospheric magnetograms and associated physical features to predict forthcoming flares. However, it is believed that flare initiation mechanisms often originate in the chromosphere and the lower corona. In this study, we employ deep learning as a purely data-driven approach to compare the predictive capabilities of chromospheric and coronal UV and EUV emissions across different wavelengths with those of photospheric line-of-sight magnetograms. Our findings indicate that individual EUV wavelengths can provide discriminatory power comparable or better to that of line-of-sight magnetograms. Moreover, we identify simple multimodal neural network architectures that consistently outperform single-input models, showing complementarity between the flare precursors that can be extracted from the distinct layers of the solar atmosphere. To mitigate potential biases from known misattributions in Active Region flare catalogs, our models are trained and evaluated using full-disk images and a comprehensive flare event catalog at the full-disk level. We introduce a deep-learning architecture suited for extracting temporal features from full-disk videos.

☆ Addressing Spectral Bias of Deep Neural Networks by Multi-Grade Deep Learning

Ronglong Fang, Yuesheng Xu

Deep neural networks (DNNs) suffer from the spectral bias, wherein DNNs typically exhibit a tendency to prioritize the learning of lower-frequency components of a function, struggling to capture its high-frequency features. This paper is to address this issue. Notice that a function having only low frequency components may be well-represented by a shallow neural network (SNN), a network having only a few layers. By observing that composition of low frequency functions can effectively approximate a high-frequency function, we propose to learn a function containing high-frequency components by composing several SNNs, each of which learns certain low-frequency information from the given data. We implement the proposed idea by exploiting the multi-grade deep learning (MGDL) model, a recently introduced model that trains a DNN incrementally, grade by grade, a current grade learning from the residue of the previous grade only an SNN composed with the SNNs trained in the preceding grades as features. We apply MGDL to synthetic, manifold, colored images, and MNIST datasets, all characterized by presence of high-frequency features. Our study reveals that MGDL excels at representing functions containing high-frequency information. Specifically, the neural networks learned in each grade adeptly capture some low-frequency information, allowing their compositions with SNNs learned in the previous grades effectively representing the high-frequency features. Our experimental results underscore the efficacy of MGDL in addressing the spectral bias inherent in DNNs. By leveraging MGDL, we offer insights into overcoming spectral bias limitation of DNNs, thereby enhancing the performance and applicability of deep learning models in tasks requiring the representation of high-frequency information. This study confirms that the proposed method offers a promising solution to address the spectral bias of DNNs.

☆ Neural Quantum Propagators for Driven-Dissipative Quantum Dynamics

Jiaji Zhang, Carlos L. Benavides-Riveros, Lipeng Chen

Describing the dynamics of strong-laser driven open quantum systems is a very challenging task that requires the solution of highly involved equations of motion. While machine learning techniques are being applied with some success to simulate the time evolution of individual quantum states, their use to approximate time-dependent operators (that can evolve various states) remains largely unexplored. In this work, we develop driven neural quantum propagators (NQP), a universal neural network framework that solves driven-dissipative quantum dynamics by approximating propagators rather than wavefunctions or density matrices. NQP can handle arbitrary initial quantum states, adapt to various external fields, and simulate long-time dynamics, even when trained on far shorter time windows. Furthermore, by appropriately configuring the external fields, our trained NQP can be transferred to systems governed by different Hamiltonians. We demonstrate the effectiveness of our approach by studying the spin-boson and the three-state transition Gamma models.

comment: 7 pages, comment are welcome!

☆ Multi-Sensor Fusion for UAV Classification Based on Feature Maps of Image and Radar Data

Nikos Sakellariou, Antonios Lalas, Konstantinos Votis, Dimitrios Tzovaras

The unique cost, flexibility, speed, and efficiency of modern UAVs make them an attractive choice in many applications in contemporary society. This, however, causes an ever-increasing number of reported malicious or accidental incidents, rendering the need for the development of UAV detection and classification mechanisms essential. We propose a methodology for developing a system that fuses already processed multi-sensor data into a new Deep Neural Network to increase its classification accuracy towards UAV detection. The DNN model fuses high-level features extracted from individual object detection and classification models associated with thermal, optronic, and radar data. Additionally, emphasis is given to the model's Convolutional Neural Network (CNN) based architecture that combines the features of the three sensor modalities by stacking the extracted image features of the thermal and optronic sensor achieving higher classification accuracy than each sensor alone.

comment: 10 pages, 6 figures

☆ Fine-Tuning LLMs for Reliable Medical Question-Answering Services ICDM

Ali Anaissi, Ali Braytee, Junaid Akram

We present an advanced approach to medical question-answering (QA) services, using fine-tuned Large Language Models (LLMs) to improve the accuracy and reliability of healthcare information. Our study focuses on optimizing models like LLaMA-2 and Mistral, which have shown great promise in delivering precise, reliable medical answers. By leveraging comprehensive datasets, we applied fine-tuning techniques such as rsDoRA+ and ReRAG. rsDoRA+ enhances model performance through a combination of decomposed model weights, varied learning rates for low-rank matrices, and rank stabilization, leading to improved efficiency. ReRAG, which integrates retrieval on demand and question rewriting, further refines the accuracy of the responses. This approach enables healthcare providers to access fast, dependable information, aiding in more efficient decision-making and fostering greater patient trust. Our work highlights the potential of fine-tuned LLMs to significantly improve the quality and accessibility of medical information services, ultimately contributing to better healthcare outcomes for all.

comment: 8 pages, 10 figures, accepted and to be published in the proceedings of 2024 IEEE International Conference on Data Mining Workshops (ICDMW)

☆ Critical Example Mining for Vehicle Trajectory Prediction using Flow-based Generative Models

Zhezhang Ding, Huijing Zhao

Precise trajectory prediction in complex driving scenarios is essential for autonomous vehicles. In practice, different driving scenarios present varying levels of difficulty for trajectory prediction models. However, most existing research focuses on the average precision of prediction results, while ignoring the underlying distribution of the input scenarios. This paper proposes a critical example mining method that utilizes a data-driven approach to estimate the rareness of the trajectories. By combining the rareness estimation of observations with whole trajectories, the proposed method effectively identifies a subset of data that is relatively hard to predict BEFORE feeding them to a specific prediction model. The experimental results show that the mined subset has higher prediction error when applied to different downstream prediction models, which reaches +108.1% error (greater than two times compared to the average on dataset) when mining 5% samples. Further analysis indicates that the mined critical examples include uncommon cases such as sudden brake and cancelled lane-change, which helps to better understand and improve the performance of prediction models.

comment: 8 pages,6 figures

☆ On-Device LLMs for SMEs: Challenges and Opportunities

Jeremy Stephen Gabriel Yee Zhi Wen, Pai Chet Ng, Zhengkui Wang, Ian McLoughlin, Aik Beng Ng, Simon See

This paper presents a systematic review of the infrastructure requirements for deploying Large Language Models (LLMs) on-device within the context of small and medium-sized enterprises (SMEs), focusing on both hardware and software perspectives. From the hardware viewpoint, we discuss the utilization of processing units like GPUs and TPUs, efficient memory and storage solutions, and strategies for effective deployment, addressing the challenges of limited computational resources typical in SME settings. From the software perspective, we explore framework compatibility, operating system optimization, and the use of specialized libraries tailored for resource-constrained environments. The review is structured to first identify the unique challenges faced by SMEs in deploying LLMs on-device, followed by an exploration of the opportunities that both hardware innovations and software adaptations offer to overcome these obstacles. Such a structured review provides practical insights, contributing significantly to the community by enhancing the technological resilience of SMEs in integrating LLMs.

comment: 9 pages, 1 figure. The work is supported by the SIT-NVIDIA Joint AI Centre

☆ Integrated Image-Text Based on Semi-supervised Learning for Small Sample Instance Segmentation

Ruting Chi, Zhiyi Huang, Yuexing Han

Small sample instance segmentation is a very challenging task, and many existing methods follow the training strategy of meta-learning which pre-train models on support set and fine-tune on query set. The pre-training phase, which is highly task related, requires a significant amount of additional training time and the selection of datasets with close proximity to ensure effectiveness. The article proposes a novel small sample instance segmentation solution from the perspective of maximizing the utilization of existing information without increasing annotation burden and training costs. The proposed method designs two modules to address the problems encountered in small sample instance segmentation. First, it helps the model fully utilize unlabeled data by learning to generate pseudo labels, increasing the number of available samples. Second, by integrating the features of text and image, more accurate classification results can be obtained. These two modules are suitable for box-free and box-dependent frameworks. In the way, the proposed method not only improves the performance of small sample instance segmentation, but also greatly reduce reliance on pre-training. We have conducted experiments in three datasets from different scenes: on land, underwater and under microscope. As evidenced by our experiments, integrated image-text corrects the confidence of classification, and pseudo labels help the model obtain preciser masks. All the results demonstrate the effectiveness and superiority of our method.

☆ TimeMixer++: A General Time Series Pattern Machine for Universal Predictive Analysis

Shiyu Wang, Jiawei Li, Xiaoming Shi, Zhou Ye, Baichuan Mo, Wenze Lin, Shengtong Ju, Zhixuan Chu, Ming Jin

Time series analysis plays a critical role in numerous applications, supporting tasks such as forecasting, classification, anomaly detection, and imputation. In this work, we present the time series pattern machine (TSPM), a model designed to excel in a broad range of time series tasks through powerful representation and pattern extraction capabilities. Traditional time series models often struggle to capture universal patterns, limiting their effectiveness across diverse tasks. To address this, we define multiple scales in the time domain and various resolutions in the frequency domain, employing various mixing strategies to extract intricate, task-adaptive time series patterns. Specifically, we introduce a general-purpose TSPM that processes multi-scale time series using (1) multi-resolution time imaging (MRTI), (2) time image decomposition (TID), (3) multi-scale mixing (MCM), and (4) multi-resolution mixing (MRM) to extract comprehensive temporal patterns. MRTI transforms multi-scale time series into multi-resolution time images, capturing patterns across both temporal and frequency domains. TID leverages dual-axis attention to extract seasonal and trend patterns, while MCM hierarchically aggregates these patterns across scales. MRM adaptively integrates all representations across resolutions. This method achieves state-of-the-art performance across 8 time series analytical tasks, consistently surpassing both general-purpose and task-specific models. Our work marks a promising step toward the next generation of TSPMs, paving the way for further advancements in time series analysis.

☆ A New Approach to Solving SMAC Task: Generating Decision Tree Code from Large Language Models

Yue Deng, Weiyu Ma, Yuxin Fan, Yin Zhang, Haifeng Zhang, Jian Zhao

StarCraft Multi-Agent Challenge (SMAC) is one of the most commonly used experimental environments in multi-agent reinforcement learning (MARL), where the specific task is to control a set number of allied units to defeat enemy forces. Traditional MARL algorithms often require interacting with the environment for up to 1 million steps to train a model, and the resulting policies are typically non-interpretable with weak transferability. In this paper, we propose a novel approach to solving SMAC tasks called LLM-SMAC. In our framework, agents leverage large language models (LLMs) to generate decision tree code by providing task descriptions. The model is further self-reflection using feedback from the rewards provided by the environment. We conduct experiments in the SMAC and demonstrate that our method can produce high-quality, interpretable decision trees with minimal environmental exploration. Moreover, these models exhibit strong transferability, successfully applying to similar SMAC environments without modification. We believe this approach offers a new direction for solving decision-making tasks in the future.

☆ Massimo: Public Queue Monitoring and Management using Mass-Spring Model

Abhijeet Kumar, Unnati Singh, Rajdeep Chatterjee, Tathagata Bandyopadhyay

An efficient system of a queue control and regulation in public spaces is very important in order to avoid the traffic jams and to improve the customer satisfaction. This article offers a detailed road map based on a merger of intelligent systems and creating an efficient systems of queues in public places. Through the utilization of different technologies i.e. computer vision, machine learning algorithms, deep learning our system provide accurate information about the place is crowded or not and the necessary efforts to be taken.

comment: 8 pages, 6 figures, 3 algorithms, 3 tables

☆ CA*: Addressing Evaluation Pitfalls in Computation-Aware Latency for Simultaneous Speech Translation

Xi Xu, Wenda Xu, Siqi Ouyang, Lei Li

Simultaneous speech translation (SimulST) systems must balance translation quality with response time, making latency measurement crucial for evaluating their real-world performance. However, there has been a longstanding belief that current metrics yield unrealistically high latency measurements in unsegmented streaming settings. In this paper, we investigate this phenomenon, revealing its root cause in a fundamental misconception underlying existing latency evaluation approaches. We demonstrate that this issue affects not only streaming but also segment-level latency evaluation across different metrics. Furthermore, we propose a modification to correctly measure computation-aware latency for SimulST systems, addressing the limitations present in existing metrics.

☆ Resilient Temporal GCN for Smart Grid State Estimation Under Topology Inaccuracies

Seyed Hamed Haghshenas, Mia Naeini

State Estimation is a crucial task in power systems. Graph Neural Networks have demonstrated significant potential in state estimation for power systems by effectively analyzing measurement data and capturing the complex interactions and interrelations among the measurements through the system's graph structure. However, the information about the system's graph structure may be inaccurate due to noise, attack or lack of accurate information about the topology of the system. This paper studies these scenarios under topology uncertainties and evaluates the impact of the topology uncertainties on the performance of a Temporal Graph Convolutional Network (TGCN) for state estimation in power systems. In order to make the model resilient to topology uncertainties, modifications in the TGCN model are proposed to incorporate a knowledge graph, generated based on the measurement data. This knowledge graph supports the assumed uncertain system graph. Two variations of the TGCN architecture are introduced to integrate the knowledge graph, and their performances are evaluated and compared to demonstrate improved resilience against topology uncertainties. The evaluation results indicate that while the two proposed architecture show different performance, they both improve the performance of the TGCN state estimation under topology uncertainties.

comment: 9 pages, 5 figures

☆ Are Language Model Logits Calibrated?

Charles Lovering, Michael Krumdick, Viet Dac Lai, Nilesh Kumar, Varshini Reddy, Rik Koncel-Kedziorski, Chris Tanner

Some information is factual (e.g., "Paris is in France"), whereas other information is probabilistic (e.g., "the coin flip will be a [Heads/Tails]."). We believe that good Language Models (LMs) should understand and reflect this nuance. Our work investigates this by testing if LMs' output probabilities are calibrated to their textual contexts. We define model "calibration" as the degree to which the output probabilities of candidate tokens are aligned with the relative likelihood that should be inferred from the given context. For example, if the context concerns two equally likely options (e.g., heads or tails for a fair coin), the output probabilities should reflect this. Likewise, context that concerns non-uniformly likely events (e.g., rolling a six with a die) should also be appropriately captured with proportionate output probabilities. We find that even in simple settings the best LMs (1) are poorly calibrated, and (2) have systematic biases (e.g., preferred colors and sensitivities to word orderings). For example, gpt-4o-mini often picks the first of two options presented in the prompt regardless of the options' implied likelihood, whereas Llama-3.1-8B picks the second. Our other consistent finding is mode-collapse: Instruction-tuned models often over-allocate probability mass on a single option. These systematic biases introduce non-intuitive model behavior, making models harder for users to understand.

comment: 10 pages (main), 24 pages (appendix), under review

☆ 1024m at SMM4H 2024: Tasks 3, 5 & 6 -- Ensembles of Transformers and Large Language Models for Medical Text Classification

Ram Mohan Rao Kadiyala, M. V. P. Chandra Sekhara Rao

Social media is a great source of data for users reporting information and regarding their health and how various things have had an effect on them. This paper presents various approaches using Transformers and Large Language Models and their ensembles, their performance along with advantages and drawbacks for various tasks of SMM4H'24 - Classifying texts on impact of nature and outdoor spaces on the author's mental health (Task 3), Binary classification of tweets reporting their children's health disorders like Asthma, Autism, ADHD and Speech disorder (task 5), Binary classification of users self-reporting their age (task 6).

comment: short paper , acl 2024

☆ Augmenting Legal Decision Support Systems with LLM-based NLI for Analyzing Social Media Evidence

Ram Mohan Rao Kadiyala, Siddartha Pullakhandam, Kanwal Mehreen, Subhasya Tippareddy, Ashay Srivastava

This paper presents our system description and error analysis of our entry for NLLP 2024 shared task on Legal Natural Language Inference (L-NLI) \citep{hagag2024legallenssharedtask2024}. The task required classifying these relationships as entailed, contradicted, or neutral, indicating any association between the review and the complaint. Our system emerged as the winning submission, significantly outperforming other entries with a substantial margin and demonstrating the effectiveness of our approach in legal text analysis. We provide a detailed analysis of the strengths and limitations of each model and approach tested, along with a thorough error analysis and suggestions for future improvements. This paper aims to contribute to the growing field of legal NLP by offering insights into advanced techniques for natural language inference in legal contexts, making it accessible to both experts and newcomers in the field.

comment: 8 pages , accepted to emnlp 2024

☆ Analyzing Closed-loop Training Techniques for Realistic Traffic Agent Models in Autonomous Highway Driving Simulations

Matthias Bitzer, Reinis Cimurs, Benjamin Coors, Johannes Goth, Sebastian Ziesche, Philipp Geiger, Maximilian Naumann

Simulation plays a crucial role in the rapid development and safe deployment of autonomous vehicles. Realistic traffic agent models are indispensable for bridging the gap between simulation and the real world. Many existing approaches for imitating human behavior are based on learning from demonstration. However, these approaches are often constrained by focusing on individual training strategies. Therefore, to foster a broader understanding of realistic traffic agent modeling, in this paper, we provide an extensive comparative analysis of different training principles, with a focus on closed-loop methods for highway driving simulation. We experimentally compare (i) open-loop vs. closed-loop multi-agent training, (ii) adversarial vs. deterministic supervised training, (iii) the impact of reinforcement losses, and (iv) the impact of training alongside log-replayed agents to identify suitable training techniques for realistic agent modeling. Furthermore, we identify promising combinations of different closed-loop training methods.

comment: 15 pages, 6 figures, 4 tables

☆ PROMPTHEUS: A Human-Centered Pipeline to Streamline SLRs with LLMs

João Pedro Fernandes Torres, Catherine Muligan, Joaquim Jorge, Catarina Moreira

The growing volume of academic publications poses significant challenges for researchers conducting timely and accurate Systematic Literature Reviews, particularly in fast-evolving fields like artificial intelligence. This growth of academic literature also makes it increasingly difficult for lay people to access scientific knowledge effectively, meaning academic literature is often misrepresented in the popular press and, more broadly, in society. Traditional SLR methods are labor-intensive and error-prone, and they struggle to keep up with the rapid pace of new research. To address these issues, we developed \textit{PROMPTHEUS}: an AI-driven pipeline solution that automates the SLR process using Large Language Models. We aimed to enhance efficiency by reducing the manual workload while maintaining the precision and coherence required for comprehensive literature synthesis. PROMPTHEUS automates key stages of the SLR process, including systematic search, data extraction, topic modeling using BERTopic, and summarization with transformer models. Evaluations conducted across five research domains demonstrate that PROMPTHEUS reduces review time, achieves high precision, and provides coherent topic organization, offering a scalable and effective solution for conducting literature reviews in an increasingly crowded research landscape. In addition, such tools may reduce the increasing mistrust in science by making summarization more accessible to laypeople. The code for this project can be found on the GitHub repository at https://github.com/joaopftorres/PROMPTHEUS.git

☆ Enabling Energy-Efficient Deployment of Large Language Models on Memristor Crossbar: A Synergy of Large and Small

Zhehui Wang, Tao Luo, Cheng Liu, Weichen Liu, Rick Siow Mong Goh, Weng-Fai Wong

Large language models (LLMs) have garnered substantial attention due to their promising applications in diverse domains. Nevertheless, the increasing size of LLMs comes with a significant surge in the computational requirements for training and deployment. Memristor crossbars have emerged as a promising solution, which demonstrated a small footprint and remarkably high energy efficiency in computer vision (CV) models. Memristors possess higher density compared to conventional memory technologies, making them highly suitable for effectively managing the extreme model size associated with LLMs. However, deploying LLMs on memristor crossbars faces three major challenges. Firstly, the size of LLMs increases rapidly, already surpassing the capabilities of state-of-the-art memristor chips. Secondly, LLMs often incorporate multi-head attention blocks, which involve non-weight stationary multiplications that traditional memristor crossbars cannot support. Third, while memristor crossbars excel at performing linear operations, they are not capable of executing complex nonlinear operations in LLM such as softmax and layer normalization. To address these challenges, we present a novel architecture for the memristor crossbar that enables the deployment of state-of-the-art LLM on a single chip or package, eliminating the energy and time inefficiencies associated with off-chip communication. Our testing on BERT_Large showed negligible accuracy loss. Compared to traditional memristor crossbars, our architecture achieves enhancements of up to 39X in area overhead and 18X in energy consumption. Compared to modern TPU/GPU systems, our architecture demonstrates at least a 68X reduction in the area-delay product and a significant 69% energy consumption reduction.

☆ Large Language Models for Cross-lingual Emotion Detection

Ram Mohan Rao Kadiyala

This paper presents a detailed system description of our entry for the WASSA 2024 Task 2, focused on cross-lingual emotion detection. We utilized a combination of large language models (LLMs) and their ensembles to effectively understand and categorize emotions across different languages. Our approach not only outperformed other submissions with a large margin, but also demonstrated the strength of integrating multiple models to enhance performance. Additionally, We conducted a thorough comparison of the benefits and limitations of each model used. An error analysis is included along with suggested areas for future improvement. This paper aims to offer a clear and comprehensive understanding of advanced techniques in emotion detection, making it accessible even to those new to the field.

comment: 6 pages , accepted to acl 2024

☆ Karush-Kuhn-Tucker Condition-Trained Neural Networks (KKT Nets)

Shreya Arvind, Rishabh Pomaje, Rajshekhar V Bhat

This paper presents a novel approach to solving convex optimization problems by leveraging the fact that, under certain regularity conditions, any set of primal or dual variables satisfying the Karush-Kuhn-Tucker (KKT) conditions is necessary and sufficient for optimality. Similar to Theory-Trained Neural Networks (TTNNs), the parameters of the convex optimization problem are input to the neural network, and the expected outputs are the optimal primal and dual variables. A choice for the loss function in this case is a loss, which we refer to as the KKT Loss, that measures how well the network's outputs satisfy the KKT conditions. We demonstrate the effectiveness of this approach using a linear program as an example. For this problem, we observe that minimizing the KKT Loss alone outperforms training the network with a weighted sum of the KKT Loss and a Data Loss (the mean-squared error between the ground truth optimal solutions and the network's output). Moreover, minimizing only the Data Loss yields inferior results compared to those obtained by minimizing the KKT Loss. While the approach is promising, the obtained primal and dual solutions are not sufficiently close to the ground truth optimal solutions. In the future, we aim to develop improved models to obtain solutions closer to the ground truth and extend the approach to other problem classes.

☆ Self-Explained Keywords Empower Large Language Models for Code Generation

Lishui Fan, Mouxiang Chen, Zhongxin Liu

Large language models (LLMs) have achieved impressive performance in code generation. However, due to the long-tail distribution of LLMs' training data, low-frequency terms are typically underrepresented in the training process. Consequently, LLMs often misunderstand or overlook problem-specific, low-frequency keywords during code generation, compromising the accuracy of the generated code. To address this, we propose a novel technique named SEK(\textbf{S}elf-\textbf{E}xplained \textbf{K}eywords), which empowers an LLM for better code generation by extracting and explaining the key terms in the problem description with the LLM itself and ranking them based on frequency. Comprehensive experiments across three benchmarks, i.e., HumanEval(+), MBPP(+), and APPS, with five representative LLMs, show that SEK can significantly improve LLMs in code generation, yielding substantial and consistent gains. For instance, SEK improves the Pass@1 of DeepSeek-Coder-V2-Instruct from 85.4\% to 93.3\% on the Humaneval benchmark. Further analysis confirms that SEK enables the LLMs to shift their attention from low-frequency keywords to their corresponding high-frequency counterparts.

☆ Systematic Exploration of Dialogue Summarization Approaches for Reproducibility, Comparative Assessment, and Methodological Innovations for Advancing Natural Language Processing in Abstractive Summarization

Yugandhar Reddy Gogireddy, Jithendra Reddy Gogireddy

Reproducibility in scientific research, particularly within the realm of natural language processing (NLP), is essential for validating and verifying the robustness of experimental findings. This paper delves into the reproduction and evaluation of dialogue summarization models, focusing specifically on the discrepancies observed between original studies and our reproduction efforts. Dialogue summarization is a critical aspect of NLP, aiming to condense conversational content into concise and informative summaries, thus aiding in efficient information retrieval and decision-making processes. Our research involved a thorough examination of several dialogue summarization models using the AMI (Augmented Multi-party Interaction) dataset. The models assessed include Hierarchical Memory Networks (HMNet) and various versions of Pointer-Generator Networks (PGN), namely PGN(DKE), PGN(DRD), PGN(DTS), and PGN(DALL). The primary objective was to evaluate the informativeness and quality of the summaries generated by these models through human assessment, a method that introduces subjectivity and variability in the evaluation process. The analysis began with Dataset 1, where the sample standard deviation of 0.656 indicated a moderate dispersion of data points around the mean.

☆ AI-Driven Innovations in Modern Cloud Computing

Animesh Kumar

The world has witnessed rapid technological transformation, past couple of decades and with Advent of Cloud computing the landscape evolved exponentially leading to efficient and scalable application development. Now, the past couple of years the digital ecosystem has brought in numerous innovations with integration of Artificial Intelligence commonly known as AI. This paper explores how AI and cloud computing intersect to deliver transformative capabilities for modernizing applications by providing services and infrastructure. Harnessing the combined potential of both AI & Cloud technologies, technology providers can now exploit intelligent resource management, predictive analytics, automated deployment & scaling with enhanced security leading to offering innovative solutions to their customers. Furthermore, by leveraging such technologies of cloud & AI businesses can reap rich rewards in the form of reducing operational costs and improving service delivery. This paper further addresses challenges associated such as data privacy concerns and how it can be mitigated with robust AI governance frameworks.

comment: 5 pages, 3 figures

☆ Do Large Language Models Have an English Accent? Evaluating and Improving the Naturalness of Multilingual LLMs

Yanzhu Guo, Simone Conia, Zelin Zhou, Min Li, Saloni Potdar, Henry Xiao

Current Large Language Models (LLMs) are predominantly designed with English as the primary language, and even the few that are multilingual tend to exhibit strong English-centric biases. Much like speakers who might produce awkward expressions when learning a second language, LLMs often generate unnatural outputs in non-English languages, reflecting English-centric patterns in both vocabulary and grammar. Despite the importance of this issue, the naturalness of multilingual LLM outputs has received limited attention. In this paper, we address this gap by introducing novel automatic corpus-level metrics to assess the lexical and syntactic naturalness of LLM outputs in a multilingual context. Using our new metrics, we evaluate state-of-the-art LLMs on a curated benchmark in French and Chinese, revealing a tendency towards English-influenced patterns. To mitigate this issue, we also propose a simple and effective alignment method to improve the naturalness of an LLM in a target language and domain, achieving consistent improvements in naturalness without compromising the performance on general-purpose benchmarks. Our work highlights the importance of developing multilingual metrics, resources and methods for the new wave of multilingual LLMs.

☆ TS-ACL: A Time Series Analytic Continual Learning Framework for Privacy-Preserving and Class-Incremental Pattern Recognition

Kejia Fan, Jiaxu Li, Songning Lai, Linpu Lv, Anfeng Liu, Jianheng Tang, Houbing Herbert Song, Huiping Zhuang

Class-incremental Learning (CIL) in Time Series Classification (TSC) aims to incrementally train models using the streaming time series data that arrives continuously. The main problem in this scenario is catastrophic forgetting, i.e., training models with new samples inevitably leads to the forgetting of previously learned knowledge. Among existing methods, the replay-based methods achieve satisfactory performance but compromise privacy, while exemplar-free methods protect privacy but suffer from low accuracy. However, more critically, owing to their reliance on gradient-based update techniques, these existing methods fundamentally cannot solve the catastrophic forgetting problem. In TSC scenarios with continuously arriving data and temporally shifting distributions, these methods become even less practical. In this paper, we propose a Time Series Analytic Continual Learning framework, called TS-ACL. Inspired by analytical learning, TS-ACL transforms neural network updates into gradient-free linear regression problems, thereby fundamentally mitigating catastrophic forgetting. Specifically, employing a pre-trained and frozen feature extraction encoder, TS-ACL only needs to update its analytic classifier recursively in a lightweight manner that is highly suitable for real-time applications and large-scale data processing. Additionally, we theoretically demonstrate that the model obtained recursively through the TS-ACL is exactly equivalent to a model trained on the complete dataset in a centralized manner, thereby establishing the property of absolute knowledge memory. Extensive experiments validate the superior performance of our TS-ACL.

comment: 11 pages, 3 figures, 2 tables

☆ User-centric evaluation of explainability of AI with and for humans: a comprehensive empirical study

Szymon Bobek, Paloma Korycińska, Monika Krakowska, Maciej Mozolewski, Dorota Rak, Magdalena Zych, Magdalena Wójcik, Grzegorz J. Nalepa

This study is located in the Human-Centered Artificial Intelligence (HCAI) and focuses on the results of a user-centered assessment of commonly used eXplainable Artificial Intelligence (XAI) algorithms, specifically investigating how humans understand and interact with the explanations provided by these algorithms. To achieve this, we employed a multi-disciplinary approach that included state-of-the-art research methods from social sciences to measure the comprehensibility of explanations generated by a state-of-the-art lachine learning model, specifically the Gradient Boosting Classifier (XGBClassifier). We conducted an extensive empirical user study involving interviews with 39 participants from three different groups, each with varying expertise in data science, data visualization, and domain-specific knowledge related to the dataset used for training the machine learning model. Participants were asked a series of questions to assess their understanding of the model's explanations. To ensure replicability, we built the model using a publicly available dataset from the UC Irvine Machine Learning Repository, focusing on edible and non-edible mushrooms. Our findings reveal limitations in existing XAI methods and confirm the need for new design principles and evaluation techniques that address the specific information needs and user perspectives of different classes of AI stakeholders. We believe that the results of our research and the cross-disciplinary methodology we developed can be successfully adapted to various data types and user profiles, thus promoting dialogue and address opportunities in HCAI research. To support this, we are making the data resulting from our study publicly available.

☆ Redefining Finance: The Influence of Artificial Intelligence (AI) and Machine Learning (ML)

Animesh Kumar

With rapid transformation of technologies, the fusion of Artificial Intelligence (AI) and Machine Learning (ML) in finance is disrupting the entire ecosystem and operations which were followed for decades. The current landscape is where decisions are increasingly data-driven by financial institutions with an appetite for automation while mitigating risks. The segments of financial institutions which are getting heavily influenced are retail banking, wealth management, corporate banking & payment ecosystem. The solution ranges from onboarding the customers all the way fraud detection & prevention to enhancing the customer services. Financial Institutes are leap frogging with integration of Artificial Intelligence and Machine Learning in mainstream applications and enhancing operational efficiency through advanced predictive analytics, extending personalized customer experiences, and automation to minimize risk with fraud detection techniques. However, with Adoption of AI & ML, it is imperative that the financial institute also needs to address ethical and regulatory challenges, by putting in place robust governance frameworks and responsible AI practices.

comment: 10 pages, 1 figure

☆ AI-Driven Approaches for Glaucoma Detection -- A Comprehensive Review

Yuki Hagiwara, Octavia-Andreaa Ciora, Maureen Monnet, Gino Lancho, Jeanette Miriam Lorenz

The diagnosis of glaucoma plays a critical role in the management and treatment of this vision-threatening disease. Glaucoma is a group of eye diseases that cause blindness by damaging the optic nerve at the back of the eye. Often called "silent thief of sight", it exhibits no symptoms during the early stages. Therefore, early detection is crucial to prevent vision loss. With the rise of Artificial Intelligence (AI), particularly Deep Learning (DL) techniques, Computer-Aided Diagnosis (CADx) systems have emerged as promising tools to assist clinicians in accurately diagnosing glaucoma early. This paper aims to provide a comprehensive overview of AI techniques utilized in CADx systems for glaucoma diagnosis. Through a detailed analysis of current literature, we identify key gaps and challenges in these systems, emphasizing the need for improved safety, reliability, interpretability, and explainability. By identifying research gaps, we aim to advance the field of CADx systems especially for the early diagnosis of glaucoma, in order to prevent any potential loss of vision.

☆ Developing Retrieval Augmented Generation (RAG) based LLM Systems from PDFs: An Experience Report

Ayman Asad Khan, Md Toufique Hasan, Kai Kristian Kemell, Jussi Rasku, Pekka Abrahamsson

This paper presents an experience report on the development of Retrieval Augmented Generation (RAG) systems using PDF documents as the primary data source. The RAG architecture combines generative capabilities of Large Language Models (LLMs) with the precision of information retrieval. This approach has the potential to redefine how we interact with and augment both structured and unstructured knowledge in generative models to enhance transparency, accuracy, and contextuality of responses. The paper details the end-to-end pipeline, from data collection, preprocessing, to retrieval indexing and response generation, highlighting technical challenges and practical solutions. We aim to offer insights to researchers and practitioners developing similar systems using two distinct approaches: OpenAI's Assistant API with GPT Series and Llama's open-source models. The practical implications of this research lie in enhancing the reliability of generative AI systems in various sectors where domain-specific knowledge and real-time information retrieval is important. The Python code used in this work is also available at: https://github.com/GPT-Laboratory/RAG-LLM-Development-Guidebook-from-PDFs.

comment: 36 pages, 8 figures, 2 tables, and python code snippets

☆ Centrality-aware Product Retrieval and Ranking EMNLP 2024

Hadeel Saadany, Swapnil Bhosale, Samarth Agrawal, Diptesh Kanojia, Constantin Orasan, Zhe Wu

This paper addresses the challenge of improving user experience on e-commerce platforms by enhancing product ranking relevant to users' search queries. Ambiguity and complexity of user queries often lead to a mismatch between the user's intent and retrieved product titles or documents. Recent approaches have proposed the use of Transformer-based models, which need millions of annotated query-title pairs during the pre-training stage, and this data often does not take user intent into account. To tackle this, we curate samples from existing datasets at eBay, manually annotated with buyer-centric relevance scores and centrality scores, which reflect how well the product title matches the users' intent. We introduce a User-intent Centrality Optimization (UCO) approach for existing models, which optimises for the user intent in semantic product search. To that end, we propose a dual-loss based optimisation to handle hard negatives, i.e., product titles that are semantically relevant but do not reflect the user's intent. Our contributions include curating challenging evaluation sets and implementing UCO, resulting in significant product ranking efficiency improvements observed for different evaluation metrics. Our work aims to ensure that the most buyer-centric titles for a query are ranked higher, thereby, enhancing the user experience on e-commerce platforms.

comment: EMNLP 2024: Industry track

☆ GReFEL: Geometry-Aware Reliable Facial Expression Learning under Bias and Imbalanced Data Distribution ACCV 2024

Azmine Toushik Wasi, Taki Hasan Rafi, Raima Islam, Karlo Serbetar, Dong Kyu Chae

Reliable facial expression learning (FEL) involves the effective learning of distinctive facial expression characteristics for more reliable, unbiased and accurate predictions in real-life settings. However, current systems struggle with FEL tasks because of the variance in people's facial expressions due to their unique facial structures, movements, tones, and demographics. Biased and imbalanced datasets compound this challenge, leading to wrong and biased prediction labels. To tackle these, we introduce GReFEL, leveraging Vision Transformers and a facial geometry-aware anchor-based reliability balancing module to combat imbalanced data distributions, bias, and uncertainty in facial expression learning. Integrating local and global data with anchors that learn different facial data points and structural features, our approach adjusts biased and mislabeled emotions caused by intra-class disparity, inter-class similarity, and scale sensitivity, resulting in comprehensive, accurate, and reliable facial expression predictions. Our model outperforms current state-of-the-art methodologies, as demonstrated by extensive experiments on various datasets.

comment: ACCV 2024. Extended version of ARBEx (arXiv:2305.01486)

☆ Bench4Merge: A Comprehensive Benchmark for Merging in Realistic Dense Traffic with Micro-Interactive Vehicles

Zhengming Wang, Junli Wang, Pengfei Li, Zhaohan Li, Peng Li, Yilun Chen

While the capabilities of autonomous driving have advanced rapidly, merging into dense traffic remains a significant challenge, many motion planning methods for this scenario have been proposed but it is hard to evaluate them. Most existing closed-loop simulators rely on rule-based controls for other vehicles, which results in a lack of diversity and randomness, thus failing to accurately assess the motion planning capabilities in highly interactive scenarios. Moreover, traditional evaluation metrics are insufficient for comprehensively evaluating the performance of merging in dense traffic. In response, we proposed a closed-loop evaluation benchmark for assessing motion planning capabilities in merging scenarios. Our approach involves other vehicles trained in large scale datasets with micro-behavioral characteristics that significantly enhance the complexity and diversity. Additionally, we have restructured the evaluation mechanism by leveraging large language models to assess each autonomous vehicle merging onto the main road. Extensive experiments have demonstrated the advanced nature of this evaluation benchmark. Through this benchmark, we have obtained an evaluation of existing methods and identified common issues. The environment and vehicle motion planning models we have designed can be accessed at https://anonymous.4open.science/r/Bench4Merge-EB5D

comment: 6 pages, 7 figures, IEEE international conference on robotics and automation

★ Diverse Policies Recovering via Pointwise Mutual Information Weighted Imitation Learning

Hanlin Yang, Jian Yao, Weiming Liu, Qing Wang, Hanmin Qin, Hansheng Kong, Kirk Tang, Jiechao Xiong, Chao Yu, Kai Li, Junliang Xing, Hongwu Chen, Juchao Zhuo, Qiang Fu, Yang Wei, Haobo Fu

Recovering a spectrum of diverse policies from a set of expert trajectories is an important research topic in imitation learning. After determining a latent style for a trajectory, previous diverse policies recovering methods usually employ a vanilla behavioral cloning learning objective conditioned on the latent style, treating each state-action pair in the trajectory with equal importance. Based on an observation that in many scenarios, behavioral styles are often highly relevant with only a subset of state-action pairs, this paper presents a new principled method in diverse polices recovery. In particular, after inferring or assigning a latent style for a trajectory, we enhance the vanilla behavioral cloning by incorporating a weighting mechanism based on pointwise mutual information. This additional weighting reflects the significance of each state-action pair's contribution to learning the style, thus allowing our method to focus on state-action pairs most representative of that style. We provide theoretical justifications for our new objective, and extensive empirical evaluations confirm the effectiveness of our method in recovering diverse policies from expert data.

comment: 18 pages, 6 figures

☆ IGMaxHS -- An Incremental MaxSAT Solver with Support for XOR Clauses

Ole Lübke

Recently, a novel, MaxSAT-based method for error correction in quantum computing has been proposed that requires both incremental MaxSAT solving capabilities and support for XOR constraints, but no dedicated MaxSAT solver fulfilling these criteria existed yet. We alleviate that and introduce IGMaxHS, which is based on the existing solvers iMaxHS and GaussMaxHS, but poses fewer restrictions on the XOR constraints than GaussMaxHS. IGMaxHS is fuzz tested with xwcnfuzz, an extension of wcnfuzz that can directly output XOR constraints. As a result, IGMaxHS is the only solver that reported neither incorrect unsatisfiability verdicts nor invalid models nor incoherent cost model combinations in a final fuzz testing comparison of all three solvers with 10000 instances. We detail the steps required for implementing Gaussian elimination on XOR constraints in CDCL SAT solvers, and extend the recently proposed re-entrant incremental MaxSAT solver application program interface to allow for incremental addition of XOR constraints. Finally, we show that IGMaxHS is capable of decoding quantum color codes through simulation with the Munich Quantum Toolkit.

comment: Presented at the 15th International Workshop on Pragmatics of SAT (PoS 2024, see https://www.pragmaticsofssat.org/2024/ )

☆ Model Mimic Attack: Knowledge Distillation for Provably Transferable Adversarial Examples

Kirill Lukyanov, Andrew Perminov, Denis Turdakov, Mikhail Pautov

The vulnerability of artificial neural networks to adversarial perturbations in the black-box setting is widely studied in the literature. The majority of attack methods to construct these perturbations suffer from an impractically large number of queries required to find an adversarial example. In this work, we focus on knowledge distillation as an approach to conduct transfer-based black-box adversarial attacks and propose an iterative training of the surrogate model on an expanding dataset. This work is the first, to our knowledge, to provide provable guarantees on the success of knowledge distillation-based attack on classification neural networks: we prove that if the student model has enough learning capabilities, the attack on the teacher model is guaranteed to be found within the finite number of distillation iterations.

☆ How to Build a Pre-trained Multimodal model for Simultaneously Chatting and Decision-making?

Zuojin Tang, Bin Hu, Chenyang Zhao, De Ma, Gang Pan, Bin Liu

Existing large pre-trained models typically map text input to text output in an end-to-end manner, such as ChatGPT, or map a segment of text input to a hierarchy of action decisions, such as OpenVLA. However, humans can simultaneously generate text and actions when receiving specific input signals. For example, a driver can make precise driving decisions while conversing with a friend in the passenger seat. Motivated by this observation, we consider the following question in this work: is it possible to construct a pre-trained model that can provide both language interaction and precise decision-making capabilities in dynamic open scenarios. We provide a definitive answer to this question by developing a new model architecture termed Visual Language Action model for Chatting and Decision Making (VLA4CD), and further demonstrating its performance in challenging autonomous driving tasks. Specifically, we leverage LoRA to fine-tune a pre-trained LLM with data of multiple modalities covering language, visual, and action. Unlike the existing LoRA operations used for LLM fine-tuning, we have designed new computational modules and training cost functions for VLA4CD. These designs enable VLA4CD to provide continuous-valued action decisions while outputting text responses. In contrast, existing LLMs can only output text responses, and current VLA models can only output action decisions. Moreover, these VLA models handle action data by discretizing and then tokenizing the discretized actions, a method unsuitable for complex decision-making tasks involving high-dimensional continuous-valued action vectors, such as autonomous driving. The experimental results on CARLA validate that: (1) our proposed model construction method is effective; (2) compared to the SOTA VLA model, VLA4CD can provide more accurate real-time decision-making while retaining the text interaction capability inherent to LLMs.

☆ Using GPT Models for Qualitative and Quantitative News Analytics in the 2024 US Presidental Election Process

Bohdan M. Pavlyshenko

The paper considers an approach of using Google Search API and GPT-4o model for qualitative and quantitative analyses of news through retrieval-augmented generation (RAG). This approach was applied to analyze news about the 2024 US presidential election process. Different news sources for different time periods have been analyzed. Quantitative scores generated by GPT model have been analyzed using Bayesian regression to derive trend lines. The distributions found for the regression parameters allow for the analysis of uncertainty in the election process. The obtained results demonstrate that using the GPT models for news analysis, one can get informative analytics and provide key insights that can be applied in further analyses of election processes.

☆ MI-VisionShot: Few-shot adaptation of vision-language models for slide-level classification of histopathological images

Pablo Meseguer, Rocío del Amor, Valery Naranjo

Vision-language supervision has made remarkable strides in learning visual representations from textual guidance. In digital pathology, vision-language models (VLM), pre-trained on curated datasets of histological image-captions, have been adapted to downstream tasks, such as region of interest classification. Zero-shot transfer for slide-level prediction has been formulated by MI-Zero, but it exhibits high variability depending on the textual prompts. Inspired by prototypical learning, we propose MI-VisionShot, a training-free adaptation method on top of VLMs to predict slide-level labels in few-shot learning scenarios. Our framework takes advantage of the excellent representation learning of VLM to create prototype-based classifiers under a multiple-instance setting by retrieving the most discriminative patches within each slide. Experimentation through different settings shows the ability of MI-VisionShot to surpass zero-shot transfer with lower variability, even in low-shot scenarios. Code coming soon at thttps://github.com/cvblab/MIVisionShot.

comment: Manuscript accepted for oral presentation at KES-InnovationInMedicine 2024 held on Madeira, Portugal

☆ FlickerFusion: Intra-trajectory Domain Generalizing Multi-Agent RL NeurIPS '24

Woosung Koh, Wonbeen Oh, Siyeol Kim, Suhin Shin, Hyeongjin Kim, Jaein Jang, Junghyun Lee, Se-Young Yun

Multi-agent reinforcement learning has demonstrated significant potential in addressing complex cooperative tasks across various real-world applications. However, existing MARL approaches often rely on the restrictive assumption that the number of entities (e.g., agents, obstacles) remains constant between training and inference. This overlooks scenarios where entities are dynamically removed or added during the inference trajectory -- a common occurrence in real-world environments like search and rescue missions and dynamic combat situations. In this paper, we tackle the challenge of intra-trajectory dynamic entity composition under zero-shot out-of-domain (OOD) generalization, where such dynamic changes cannot be anticipated beforehand. Our empirical studies reveal that existing MARL methods suffer significant performance degradation and increased uncertainty in these scenarios. In response, we propose FlickerFusion, a novel OOD generalization method that acts as a universally applicable augmentation technique for MARL backbone methods. Our results show that FlickerFusion not only achieves superior inference rewards but also uniquely reduces uncertainty vis-\`a-vis the backbone, compared to existing methods. For standardized evaluation, we introduce MPEv2, an enhanced version of Multi Particle Environments (MPE), consisting of 12 benchmarks. Benchmarks, implementations, and trained models are organized and open-sourced at flickerfusion305.github.io, accompanied by ample demo video renderings.

comment: NeurIPS '24 Open-World Agents Workshop

☆ Mesa-Extrapolation: A Weave Position Encoding Method for Enhanced Extrapolation in LLMs NeurIPS 2024

Xin Ma, Yang Liu, Jingjing Liu, Xiaoxu Ma

Large language models (LLMs), although having revolutionized many fields, still suffer from the challenging extrapolation problem, where the inference ability of LLMs sharply declines beyond their max training lengths. In this work, we conduct a theoretical analysis to better understand why No Position Encoding (NoPE) fails outside its effective range, as well as examining the power of Position Encoding (PE) in this context. Our findings reveal that with meticulous weave position, PE can indeed be extended beyond effective range. Our theorems establish that LLMs equipped with weave PE can achieve improved extrapolation performance without additional cost. Furthermore, we introduce a novel weave PE method, Mesa-Extrapolation, which utilizes a chunk-based triangular attention matrix and applies Stair PE to manage the final chunk. This method not only retains competitive performance but also offers substantial benefits such as significantly reduced memory demand and faster inference speed. Extensive experiments validate the effectiveness of Mesa-Extrapolation, demonstrating its potential as a scalable solution to enhancing LLMs applicative reach.

comment: accepted by NeurIPS 2024. arXiv admin note: text overlap with arXiv:2305.19466 by other authors

☆ Random Token Fusion for Multi-View Medical Diagnosis NeurIPS 2024

Jingyu Guo, Christos Matsoukas, Fredrik Strand, Kevin Smith

In multi-view medical diagnosis, deep learning-based models often fuse information from different imaging perspectives to improve diagnostic performance. However, existing approaches are prone to overfitting and rely heavily on view-specific features, which can lead to trivial solutions. In this work, we introduce Random Token Fusion (RTF), a novel technique designed to enhance multi-view medical image analysis using vision transformers. By integrating randomness into the feature fusion process during training, RTF addresses the issue of overfitting and enhances the robustness and accuracy of diagnostic models without incurring any additional cost at inference. We validate our approach on standard mammography and chest X-ray benchmark datasets. Through extensive experiments, we demonstrate that RTF consistently improves the performance of existing fusion methods, paving the way for a new generation of multi-view medical foundation models.

comment: Originally published at the NeurIPS 2024 Workshop on Advancements In Medical Foundation Models: Explainability, Robustness, Security, and Beyond (AIM-FM)

☆ Long-distance Geomagnetic Navigation in GNSS-denied Environments with Deep Reinforcement Learning

Wenqi Bai, Xiaohui Zhang, Shiliang Zhang, Songnan Yang, Yushuai Li, Tingwen Huang

Geomagnetic navigation has drawn increasing attention with its capacity in navigating through complex environments and its independence from external navigation services like global navigation satellite systems (GNSS). Existing studies on geomagnetic navigation, i.e., matching navigation and bionic navigation, rely on pre-stored map or extensive searches, leading to limited applicability or reduced navigation efficiency in unexplored areas. To address the issues with geomagnetic navigation in areas where GNSS is unavailable, this paper develops a deep reinforcement learning (DRL)-based mechanism, especially for long-distance geomagnetic navigation. The designed mechanism trains an agent to learn and gain the magnetoreception capacity for geomagnetic navigation, rather than using any pre-stored map or extensive and expensive searching approaches. Particularly, we integrate the geomagnetic gradient-based parallel approach into geomagnetic navigation. This integration mitigates the over-exploration of the learning agent by adjusting the geomagnetic gradient, such that the obtained gradient is aligned towards the destination. We explore the effectiveness of the proposed approach via detailed numerical simulations, where we implement twin delayed deep deterministic policy gradient (TD3) in realizing the proposed approach. The results demonstrate that our approach outperforms existing metaheuristic and bionic navigation methods in long-distance missions under diverse navigation conditions.

☆ LLM4GRN: Discovering Causal Gene Regulatory Networks with LLMs -- Evaluation through Synthetic Data Generation

Tejumade Afonja, Ivaxi Sheth, Ruta Binkyte, Waqar Hanif, Thomas Ulas, Matthias Becker, Mario Fritz

Gene regulatory networks (GRNs) represent the causal relationships between transcription factors (TFs) and target genes in single-cell RNA sequencing (scRNA-seq) data. Understanding these networks is crucial for uncovering disease mechanisms and identifying therapeutic targets. In this work, we investigate the potential of large language models (LLMs) for GRN discovery, leveraging their learned biological knowledge alone or in combination with traditional statistical methods. We develop a task-based evaluation strategy to address the challenge of unavailable ground truth causal graphs. Specifically, we use the GRNs suggested by LLMs to guide causal synthetic data generation and compare the resulting data against the original dataset. Our statistical and biological assessments show that LLMs can support statistical modeling and data synthesis for biological research.

☆ The effect of fine-tuning on language model toxicity NeurIPS 2024

Will Hawkins, Brent Mittelstadt, Chris Russell

Fine-tuning language models has become increasingly popular following the proliferation of open models and improvements in cost-effective parameter efficient fine-tuning. However, fine-tuning can influence model properties such as safety. We assess how fine-tuning can impact different open models' propensity to output toxic content. We assess the impacts of fine-tuning Gemma, Llama, and Phi models on toxicity through three experiments. We compare how toxicity is reduced by model developers during instruction-tuning. We show that small amounts of parameter-efficient fine-tuning on developer-tuned models via low-rank adaptation on a non-adversarial dataset can significantly alter these results across models. Finally, we highlight the impact of this in the wild, demonstrating how toxicity rates of models fine-tuned by community contributors can deviate in hard-to-predict ways.

comment: To be presented at NeurIPS 2024 Safe Generative AI Workshop

☆ MAC Revivo: Artificial Intelligence Paves the Way

Jinzhe Pan, Jingqing Wang, Zelin Yun, Zhiyong Xiao, Yuehui Ouyang, Wenchi Cheng, Wei Zhang

The vast adoption of Wi-Fi and/or Bluetooth capabilities in Internet of Things (IoT) devices, along with the rapid growth of deployed smart devices, has caused significant interference and congestion in the industrial, scientific, and medical (ISM) bands. Traditional Wi-Fi Medium Access Control (MAC) design faces significant challenges in managing increasingly complex wireless environments while ensuring network Quality of Service (QoS) performance. This paper explores the potential integration of advanced Artificial Intelligence (AI) methods into the design of Wi-Fi MAC protocols. We propose AI-MAC, an innovative approach that employs machine learning algorithms to dynamically adapt to changing network conditions, optimize channel access, mitigate interference, and ensure deterministic latency. By intelligently predicting and managing interference, AI-MAC aims to provide a robust solution for next generation of Wi-Fi networks, enabling seamless connectivity and enhanced QoS. Our experimental results demonstrate that AI-MAC significantly reduces both interference and latency, paving the way for more reliable and efficient wireless communications in the increasingly crowded ISM band.

☆ LiMTR: Time Series Motion Prediction for Diverse Road Users through Multimodal Feature Integration NeurIPS 2024

Camiel Oerlemans, Bram Grooten, Michiel Braat, Alaa Alassi, Emilia Silvas, Decebal Constantin Mocanu

Predicting the behavior of road users accurately is crucial to enable the safe operation of autonomous vehicles in urban or densely populated areas. Therefore, there has been a growing interest in time series motion prediction research, leading to significant advancements in state-of-the-art techniques in recent years. However, the potential of using LiDAR data to capture more detailed local features, such as a person's gaze or posture, remains largely unexplored. To address this, we develop a novel multimodal approach for motion prediction based on the PointNet foundation model architecture, incorporating local LiDAR features. Evaluation on the Waymo Open Dataset shows a performance improvement of 6.20% and 1.58% in minADE and mAP respectively, when integrated and compared with the previous state-of-the-art MTR. We open-source the code of our LiMTR model.

comment: Accepted at the NeurIPS 2024 workshop Time Series in the Age of Large Models. Code available at https://github.com/Cing2/LiMTR

☆ Kaninfradet3D:A Road-side Camera-LiDAR Fusion 3D Perception Model based on Nonlinear Feature Extraction and Intrinsic Correlation

Pei Liu, Nanfang Zheng, Yiqun Li, Junlan Chen, Ziyuan Pu

With the development of AI-assisted driving, numerous methods have emerged for ego-vehicle 3D perception tasks, but there has been limited research on roadside perception. With its ability to provide a global view and a broader sensing range, the roadside perspective is worth developing. LiDAR provides precise three-dimensional spatial information, while cameras offer semantic information. These two modalities are complementary in 3D detection. However, adding camera data does not increase accuracy in some studies since the information extraction and fusion procedure is not sufficiently reliable. Recently, Kolmogorov-Arnold Networks (KANs) have been proposed as replacements for MLPs, which are better suited for high-dimensional, complex data. Both the camera and the LiDAR provide high-dimensional information, and employing KANs should enhance the extraction of valuable features to produce better fusion outcomes. This paper proposes Kaninfradet3D, which optimizes the feature extraction and fusion modules. To extract features from complex high-dimensional data, the model's encoder and fuser modules were improved using KAN Layers. Cross-attention was applied to enhance feature fusion, and visual comparisons verified that camera features were more evenly integrated. This addressed the issue of camera features being abnormally concentrated, negatively impacting fusion. Compared to the benchmark, our approach shows improvements of +9.87 mAP and +10.64 mAP in the two viewpoints of the TUMTraf Intersection Dataset and an improvement of +1.40 mAP in the roadside end of the TUMTraf V2X Cooperative Perception Dataset. The results indicate that Kaninfradet3D can effectively fuse features, demonstrating the potential of applying KANs in roadside perception tasks.

☆ RAG4ITOps: A Supervised Fine-Tunable and Comprehensive RAG Framework for IT Operations and Maintenance EMNLP 2024

Tianyang Zhang, Zhuoxuan Jiang, Shengguang Bai, Tianrui Zhang, Lin Lin, Yang Liu, Jiawei Ren

With the ever-increasing demands on Question Answering (QA) systems for IT operations and maintenance, an efficient and supervised fine-tunable framework is necessary to ensure the data security, private deployment and continuous upgrading. Although Large Language Models (LLMs) have notably improved the open-domain QA's performance, how to efficiently handle enterprise-exclusive corpora and build domain-specific QA systems are still less-studied for industrial applications. In this paper, we propose a general and comprehensive framework based on Retrieval Augmented Generation (RAG) and facilitate the whole business process of establishing QA systems for IT operations and maintenance. In accordance with the prevailing RAG method, our proposed framework, named with RAG4ITOps, composes of two major stages: (1) Models Fine-tuning \& Data Vectorization, and (2) Online QA System Process. At the Stage 1, we leverage a contrastive learning method with two negative sampling strategies to fine-tune the embedding model, and design the instruction templates to fine-tune the LLM with a Retrieval Augmented Fine-Tuning method. At the Stage 2, an efficient process of QA system is built for serving. We collect enterprise-exclusive corpora from the domain of cloud computing, and the extensive experiments show that our method achieves superior results than counterparts on two kinds of QA tasks. Our experiment also provide a case for applying the RAG4ITOps to real-world enterprise-level applications.

comment: Accepted by EMNLP 2024 Industry Track

☆ Deep Learning and Data Augmentation for Detecting Self-Admitted Technical Debt

Edi Sutoyo, Paris Avgeriou, Andrea Capiluppi

Self-Admitted Technical Debt (SATD) refers to circumstances where developers use textual artifacts to explain why the existing implementation is not optimal. Past research in detecting SATD has focused on either identifying SATD (classifying SATD items as SATD or not) or categorizing SATD (labeling instances as SATD that pertain to requirement, design, code, test debt, etc.). However, the performance of these approaches remains suboptimal, particularly for specific types of SATD, such as test and requirement debt, primarily due to extremely imbalanced datasets. To address these challenges, we build on earlier research by utilizing BiLSTM architecture for the binary identification of SATD and BERT architecture for categorizing different types of SATD. Despite their effectiveness, both architectures struggle with imbalanced data. Therefore, we employ a large language model data augmentation strategy to mitigate this issue. Furthermore, we introduce a two-step approach to identify and categorize SATD across various datasets derived from different artifacts. Our contributions include providing a balanced dataset for future SATD researchers and demonstrating that our approach significantly improves SATD identification and categorization performance compared to baseline methods.

comment: Accepted to be published at the 2024 31st Asia-Pacific Software Engineering Conference (APSEC)

☆ Habaek: High-performance water segmentation through dataset expansion and inductive bias optimization

Hanseon Joo, Eunji Lee, Minjong Cheon

Water segmentation is critical to disaster response and water resource management. Authorities may employ high-resolution photography to monitor rivers, lakes, and reservoirs, allowing for more proactive management in agriculture, industry, and conservation. Deep learning has improved flood monitoring by allowing models like CNNs, U-Nets, and transformers to handle large volumes of satellite and aerial data. However, these models usually have significant processing requirements, limiting their usage in real-time applications. This research proposes upgrading the SegFormer model for water segmentation by data augmentation with datasets such as ADE20K and RIWA to boost generalization. We examine how inductive bias affects attention-based models and discover that SegFormer performs better on bigger datasets. To further demonstrate the function of data augmentation, Low-Rank Adaptation (LoRA) is used to lower processing complexity while preserving accuracy. We show that the suggested Habaek model outperforms current models in segmentation, with an Intersection over Union (IoU) ranging from 0.91986 to 0.94397. In terms of F1-score, recall, accuracy, and precision, Habaek performs better than rival models, indicating its potential for real-world applications. This study highlights the need to enhance structures and include datasets for effective water segmentation.

☆ WildOcc: A Benchmark for Off-Road 3D Semantic Occupancy Prediction

Heng Zhai, Jilin Mei, Chen Min, Liang Chen, Fangzhou Zhao, Yu Hu

3D semantic occupancy prediction is an essential part of autonomous driving, focusing on capturing the geometric details of scenes. Off-road environments are rich in geometric information, therefore it is suitable for 3D semantic occupancy prediction tasks to reconstruct such scenes. However, most of researches concentrate on on-road environments, and few methods are designed for off-road 3D semantic occupancy prediction due to the lack of relevant datasets and benchmarks. In response to this gap, we introduce WildOcc, to our knowledge, the first benchmark to provide dense occupancy annotations for off-road 3D semantic occupancy prediction tasks. A ground truth generation pipeline is proposed in this paper, which employs a coarse-to-fine reconstruction to achieve a more realistic result. Moreover, we introduce a multi-modal 3D semantic occupancy prediction framework, which fuses spatio-temporal information from multi-frame images and point clouds at voxel level. In addition, a cross-modality distillation function is introduced, which transfers geometric knowledge from point clouds to image features.

☆ Arithmetic Transformers Can Length-Generalize in Both Operand Length and Count

Hanseul Cho, Jaeyoung Cha, Srinadh Bhojanapalli, Chulhee Yun

Transformers often struggle with length generalization, meaning they fail to generalize to sequences longer than those encountered during training. While arithmetic tasks are commonly used to study length generalization, certain tasks are considered notoriously difficult, e.g., multi-operand addition (requiring generalization over both the number of operands and their lengths) and multiplication (requiring generalization over both operand lengths). In this work, we achieve approximately 2-3x length generalization on both tasks, which is the first such achievement in arithmetic Transformers. We design task-specific scratchpads enabling the model to focus on a fixed number of tokens per each next-token prediction step, and apply multi-level versions of Position Coupling (Cho et al., 2024; McLeish et al., 2024) to let Transformers know the right position to attend to. On the theory side, we prove that a 1-layer Transformer using our method can solve multi-operand addition, up to operand length and operand count that are exponential in embedding dimension.

comment: 38 pages, 16 figures

☆ An Efficient System for Automatic Map Storytelling -- A Case Study on Historical Maps

Ziyi Liu, Claudio Affolter, Sidi Wu, Yizi Chen, Lorenz Hurni

Historical maps provide valuable information and knowledge about the past. However, as they often feature non-standard projections, hand-drawn styles, and artistic elements, it is challenging for non-experts to identify and interpret them. While existing image captioning methods have achieved remarkable success on natural images, their performance on maps is suboptimal as maps are underrepresented in their pre-training process. Despite the recent advance of GPT-4 in text recognition and map captioning, it still has a limited understanding of maps, as its performance wanes when texts (e.g., titles and legends) in maps are missing or inaccurate. Besides, it is inefficient or even impractical to fine-tune the model with users' own datasets. To address these problems, we propose a novel and lightweight map-captioning counterpart. Specifically, we fine-tune the state-of-the-art vision-language model CLIP to generate captions relevant to historical maps and enrich the captions with GPT-3.5 to tell a brief story regarding where, what, when and why of a given map. We propose a novel decision tree architecture to only generate captions relevant to the specified map type. Our system shows invariance to text alterations in maps. The system can be easily adapted and extended to other map types and scaled to a larger map captioning system. The code is open-sourced at https://github.com/claudaff/automatic-map-storytelling.

☆ Reducing Hallucinations in Vision-Language Models via Latent Space Steering

Sheng Liu, Haotian Ye, James Zou

Hallucination poses a challenge to the deployment of large vision-language models (LVLMs) in applications. Unlike in large language models (LLMs), hallucination in LVLMs often arises from misalignments between visual inputs and textual outputs. This paper investigates the underlying mechanisms of hallucination, focusing on the unique structure of LVLMs that distinguishes them from large language models (LLMs). We identify that hallucinations often arise from the sensitivity of text decoders to vision inputs, a natural phenomenon when image encoders and text decoders are pre-trained separately. Inspired by this, we introduce Visual and Textual Intervention (VTI), a novel technique designed to reduce hallucinations by steering latent space representations during inference to enhance the stability of vision features. As a task-agnostic test-time intervention, VTI can be easily applied to any problem without additional cost. Extensive experiments demonstrate that it can effectively reduce hallucinations and outperform baseline methods across multiple metrics, highlighting the critical role of vision feature stability in LVLMs.

comment: 21 pages

☆ A roadmap for generative mapping: unlocking the power of generative AI for map-making

Sidi Wu, Katharina Henggeler, Yizi Chen, Lorenz Hurni

Maps are broadly relevant across various fields, serving as valuable tools for presenting spatial phenomena and communicating spatial knowledge. However, map-making is still largely confined to those with expertise in GIS and cartography due to the specialized software and complex workflow involved, from data processing to visualization. While generative AI has recently demonstrated its remarkable capability in creating various types of content and its wide accessibility to the general public, its potential in generating maps is yet to be fully realized. This paper highlights the key applications of generative AI in map-making, summarizes recent advancements in generative AI, identifies the specific technologies required and the challenges of using current methods, and provides a roadmap for developing a generative mapping system (GMS) to make map-making more accessible.

☆ Learning to Synthesize Graphics Programs for Geometric Artworks ICPR 2024

Qi Bing, Chaoyi Zhang, Weidong Cai

Creating and understanding art has long been a hallmark of human ability. When presented with finished digital artwork, professional graphic artists can intuitively deconstruct and replicate it using various drawing tools, such as the line tool, paint bucket, and layer features, including opacity and blending modes. While most recent research in this field has focused on art generation, proposing a range of methods, these often rely on the concept of artwork being represented as a final image. To bridge the gap between pixel-level results and the actual drawing process, we present an approach that treats a set of drawing tools as executable programs. This method predicts a sequence of steps to achieve the final image, allowing for understandable and resolution-independent reproductions under the usage of a set of drawing commands. Our experiments demonstrate that our program synthesizer, Art2Prog, can comprehensively understand complex input images and reproduce them using high-quality executable programs. The experimental results evidence the potential of machines to grasp higher-level information from images and generate compact program-level descriptions.

comment: ICPR 2024

☆ LSCodec: Low-Bitrate and Speaker-Decoupled Discrete Speech Codec ICASSP 2025

Yiwei Guo, Zhihan Li, Chenpeng Du, Hankun Wang, Xie Chen, Kai Yu

Although discrete speech tokens have exhibited strong potential for language model-based speech generation, their high bitrates and redundant timbre information restrict the development of such models. In this work, we propose LSCodec, a discrete speech codec that has both low bitrate and speaker decoupling ability. LSCodec adopts a three-stage unsupervised training framework with a speaker perturbation technique. A continuous information bottleneck is first established, followed by vector quantization that produces a discrete speaker-decoupled space. A discrete token vocoder finally refines acoustic details from LSCodec. By reconstruction experiments, LSCodec demonstrates superior intelligibility and audio quality with only a single codebook and smaller vocabulary size than baselines. The 25Hz version of LSCodec also achieves the lowest bitrate (0.25kbps) of codecs so far with decent quality. Voice conversion evaluations prove the satisfactory speaker disentanglement of LSCodec, and ablation study further verifies the effectiveness of the proposed training framework.

comment: 5 pages, 2 figures, 4 tables. Submitted to ICASSP 2025. Demo page: https://cantabile-kwok.github.io/LSCodec/

☆ DeepIcon: A Hierarchical Network for Layer-wise Icon Vectorization

Qi Bing, Chaoyi Zhang, Weidong Cai

In contrast to the well-established technique of rasterization, vectorization of images poses a significant challenge in the field of computer graphics. Recent learning-based methods for converting raster images to vector formats frequently suffer from incomplete shapes, redundant path prediction, and a lack of accuracy in preserving the semantics of the original content. These shortcomings severely hinder the utility of these methods for further editing and manipulation of images. To address these challenges, we present DeepIcon, a novel hierarchical image vectorization network specifically tailored for generating variable-length icon vector graphics based on the raster image input. Our experimental results indicate that DeepIcon can efficiently produce Scalable Vector Graphics (SVGs) directly from raster images, bypassing the need for a differentiable rasterizer while also demonstrating a profound understanding of the image contents.

comment: Accepted as Oral Presentation at DICTA 2024

☆ Automated Proof Generation for Rust Code via Self-Evolution

Tianyu Chen, Shuai Lu, Shan Lu, Yeyun Gong, Chenyuan Yang, Xuheng Li, Md Rakib Hossain Misu, Hao Yu, Nan Duan, Peng Cheng, Fan Yang, Shuvendu K Lahiri, Tao Xie, Lidong Zhou

Ensuring correctness is crucial for code generation. Formal verification offers a definitive assurance of correctness, but demands substantial human effort in proof construction and hence raises a pressing need for automation. The primary obstacle lies in the severe lack of data - there is much less proof than code for LLMs to train upon. In this paper, we introduce SAFE, a novel framework that overcomes the lack of human-written proof to enable automated proof generation of Rust code. SAFE establishes a self-evolving cycle where data synthesis and fine-tuning collaborate to enhance the model capability, leveraging the definitive power of a symbolic verifier in telling correct proof from incorrect ones. SAFE also re-purposes the large number of synthesized incorrect proofs to train the self-debugging capability of the fine-tuned models, empowering them to fix incorrect proofs based on the verifier's feedback. SAFE demonstrates superior efficiency and precision compared to GPT-4o. Through tens of thousands of synthesized proofs and the self-debugging mechanism, we improve the capability of open-source models, initially unacquainted with formal verification, to automatically write proof for Rust code. This advancement leads to a significant improvement in performance, achieving a 70.50% accuracy rate in a benchmark crafted by human experts, a significant leap over GPT-4o's performance of 24.46%.

☆ GIG: Graph Data Imputation With Graph Differential Dependencies

Jiang Hua, Michael Bewong, Selasi Kwashie, MD Geaur Rahman, Junwei Hu, Xi Guo, Zaiwen Fen

Data imputation addresses the challenge of imputing missing values in database instances, ensuring consistency with the overall semantics of the dataset. Although several heuristics which rely on statistical methods, and ad-hoc rules have been proposed. These do not generalise well and often lack data context. Consequently, they also lack explainability. The existing techniques also mostly focus on the relational data context making them unsuitable for wider application contexts such as in graph data. In this paper, we propose a graph data imputation approach called GIG which relies on graph differential dependencies (GDDs). GIG, learns the GDDs from a given knowledge graph, and uses these rules to train a transformer model which then predicts the value of missing data within the graph. By leveraging GDDs, GIG incoporates semantic knowledge into the data imputation process making it more reliable and explainable. Experimental results on seven real-world datasets highlight GIG's effectiveness compared to existing state-of-the-art approaches.

comment: 12 pages, 4 figures, published to ADC

☆ Alchemy: Amplifying Theorem-Proving Capability through Symbolic Mutation

Shaonan Wu, Shuai Lu, Yeyun Gong, Nan Duan, Ping Wei

Formal proofs are challenging to write even for experienced experts. Recent progress in Neural Theorem Proving (NTP) shows promise in expediting this process. However, the formal corpora available on the Internet are limited compared to the general text, posing a significant data scarcity challenge for NTP. To address this issue, this work proposes Alchemy, a general framework for data synthesis that constructs formal theorems through symbolic mutation. Specifically, for each candidate theorem in Mathlib, we identify all invocable theorems that can be used to rewrite or apply to it. Subsequently, we mutate the candidate theorem by replacing the corresponding term in the statement with its equivalent form or antecedent. As a result, our method increases the number of theorems in Mathlib by an order of magnitude, from 110k to 6M. Furthermore, we perform continual pretraining and supervised finetuning on this augmented corpus for large language models. Experimental results demonstrate the effectiveness of our approach, achieving a 5% absolute performance improvement on Leandojo benchmark. Additionally, our synthetic data achieve a 2.5% absolute performance gain on the out-of-distribution miniF2F benchmark. To provide further insights, we conduct a comprehensive analysis of synthetic data composition and the training paradigm, offering valuable guidance for developing a strong theorem prover.

☆ Unleashing the Potential of Vision-Language Pre-Training for 3D Zero-Shot Lesion Segmentation via Mask-Attribute Alignment

Yankai Jiang, Wenhui Lei, Xiaofan Zhang, Shaoting Zhang

Recent advancements in medical vision-language pre-training models have driven significant progress in zero-shot disease recognition. However, transferring image-level knowledge to pixel-level tasks, such as lesion segmentation in 3D CT scans, remains a critical challenge. Due to the complexity and variability of pathological visual characteristics, existing methods struggle to align fine-grained lesion features not encountered during training with disease-related textual representations. In this paper, we present Malenia, a novel multi-scale lesion-level mask-attribute alignment framework, specifically designed for 3D zero-shot lesion segmentation. Malenia improves the compatibility between mask representations and their associated elemental attributes, explicitly linking the visual features of unseen lesions with the extensible knowledge learned from previously seen ones. Furthermore, we design a Cross-Modal Knowledge Injection module to enhance both visual and textual features with mutually beneficial information, effectively guiding the generation of segmentation results. Comprehensive experiments across three datasets and 12 lesion categories validate the superior performance of Malenia. Codes will be publicly available.

☆ Who's Who: Large Language Models Meet Knowledge Conflicts in Practice EMNLP 2024

Quang Hieu Pham, Hoang Ngo, Anh Tuan Luu, Dat Quoc Nguyen

Retrieval-augmented generation (RAG) methods are viable solutions for addressing the static memory limits of pre-trained language models. Nevertheless, encountering conflicting sources of information within the retrieval context is an inevitable practical challenge. In such situations, the language models are recommended to transparently inform users about the conflicts rather than autonomously deciding what to present based on their inherent biases. To analyze how current large language models (LLMs) align with our recommendation, we introduce WhoQA, a public benchmark dataset to examine model's behavior in knowledge conflict situations. We induce conflicts by asking about a common property among entities having the same name, resulting in questions with up to 8 distinctive answers. WhoQA evaluation set includes 5K questions across 13 Wikidata property types and 150K Wikipedia entities. Our experiments show that despite the simplicity of WhoQA questions, knowledge conflicts significantly degrades LLMs' performance in RAG settings.

comment: Accepted to EMNLP 2024 Findings

☆ AutoTrain: No-code training for state-of-the-art models

Abhishek Thakur

With the advancements in open-source models, training (or finetuning) models on custom datasets has become a crucial part of developing solutions which are tailored to specific industrial or open-source applications. Yet, there is no single tool which simplifies the process of training across different types of modalities or tasks. We introduce AutoTrain (aka AutoTrain Advanced) -- an open-source, no code tool/library which can be used to train (or finetune) models for different kinds of tasks such as: large language model (LLM) finetuning, text classification/regression, token classification, sequence-to-sequence task, finetuning of sentence transformers, visual language model (VLM) finetuning, image classification/regression and even classification and regression tasks on tabular data. AutoTrain Advanced is an open-source library providing best practices for training models on custom datasets. The library is available at https://github.com/huggingface/autotrain-advanced. AutoTrain can be used in fully local mode or on cloud machines and works with tens of thousands of models shared on Hugging Face Hub and their variations.

☆ Reducing annotator bias by belief elicitation

Terne Sasha Thorn Jakobsen, Andreas Bjerre-Nielsen, Robert Böhm

Crowdsourced annotations of data play a substantial role in the development of Artificial Intelligence (AI). It is broadly recognised that annotations of text data can contain annotator bias, where systematic disagreement in annotations can be traced back to differences in the annotators' backgrounds. Being unaware of such annotator bias can lead to representational bias against minority group perspectives and therefore several methods have been proposed for recognising bias or preserving perspectives. These methods typically require either a substantial number of annotators or annotations per data instance. In this study, we propose a simple method for handling bias in annotations without requirements on the number of annotators or instances. Instead, we ask annotators about their beliefs of other annotators' judgements of an instance, under the hypothesis that these beliefs may provide more representative and less biased labels than judgements. The method was examined in two controlled, survey-based experiments involving Democrats and Republicans (n=1,590) asked to judge statements as arguments and then report beliefs about others' judgements. The results indicate that bias, defined as systematic differences between the two groups of annotators, is consistently reduced when asking for beliefs instead of judgements. Our proposed method therefore has the potential to reduce the risk of annotator bias, thereby improving the generalisability of AI systems and preventing harm to unrepresented socio-demographic groups, and we highlight the need for further studies of this potential in other tasks and downstream applications.

☆ Timetable Nodes for Public Transport Network

Andrii Rohovyi, Peter J. Stuckey, Toby Walsh

Faster pathfinding in time-dependent transport networks is an important and challenging problem in navigation systems. There are two main types of transport networks: road networks for car driving and public transport route network. The solutions that work well in road networks, such as Time-dependent Contraction Hierarchies and other graph-based approaches, do not usually apply in transport networks. In transport networks, non-graph solutions such as CSA and RAPTOR show the best results compared to graph-based techniques. In our work, we propose a method that advances graph-based approaches by using different optimization techniques from computational geometry to speed up the search process in transport networks. We apply a new pre-computation step, which we call timetable nodes (TTN). Our inspiration comes from an iterative search problem in computational geometry. We implement two versions of the TTN: one uses a Combined Search Tree (TTN-CST), and the second uses Fractional Cascading (TTN-FC). Both of these approaches decrease the asymptotic complexity of reaching new nodes from $O(k\times \log|C|)$ to $O(k + \log(k) + \log(|C|))$, where $k$ is the number of outgoing edges from a node and $|C|$ is the size of the timetable information (total outgoing edges). Our solution suits any other time-dependent networks and can be integrated into other pathfinding algorithms. Our experiments indicate that this pre-computation significantly enhances the performance on high-density graphs. This study showcases how leveraging computational geometry can enhance pathfinding in transport networks, enabling faster pathfinding in scenarios involving large numbers of outgoing edges.

☆ Offline reinforcement learning for job-shop scheduling problems

Imanol Echeverria, Maialen Murua, Roberto Santana

Recent advances in deep learning have shown significant potential for solving combinatorial optimization problems in real-time. Unlike traditional methods, deep learning can generate high-quality solutions efficiently, which is crucial for applications like routing and scheduling. However, existing approaches like deep reinforcement learning (RL) and behavioral cloning have notable limitations, with deep RL suffering from slow learning and behavioral cloning relying solely on expert actions, which can lead to generalization issues and neglect of the optimization objective. This paper introduces a novel offline RL method designed for combinatorial optimization problems with complex constraints, where the state is represented as a heterogeneous graph and the action space is variable. Our approach encodes actions in edge attributes and balances expected rewards with the imitation of expert solutions. We demonstrate the effectiveness of this method on job-shop scheduling and flexible job-shop scheduling benchmarks, achieving superior performance compared to state-of-the-art techniques.

☆ InternLM2.5-StepProver: Advancing Automated Theorem Proving via Expert Iteration on Large-Scale LEAN Problems

Zijian Wu, Suozhi Huang, Zhejian Zhou, Huaiyuan Ying, Jiayu Wang, Dahua Lin, Kai Chen

Large Language Models (LLMs) have emerged as powerful tools in mathematical theorem proving, particularly when utilizing formal languages such as LEAN. The major learning paradigm is expert iteration, which necessitates a pre-defined dataset comprising numerous mathematical problems. In this process, LLMs attempt to prove problems within the dataset and iteratively refine their capabilities through self-training on the proofs they discover. We propose to use large scale LEAN problem datasets Lean-workbook for expert iteration with more than 20,000 CPU days. During expert iteration, we found log-linear trends between solved problem amount with proof length and CPU usage. We train a critic model to select relatively easy problems for policy models to make trials and guide the model to search for deeper proofs. InternLM2.5-StepProver achieves open-source state-of-the-art on MiniF2F, Lean-Workbook-Plus, ProofNet, and Putnam benchmarks. Specifically, it achieves a pass of 65.9% on the MiniF2F-test and proves (or disproves) 17.0% of problems in Lean-Workbook-Plus which shows a significant improvement compared to only 9.5% of problems proved when Lean-Workbook-Plus was released. We open-source our models and searched proofs at https://github.com/InternLM/InternLM-Math and https://huggingface.co/datasets/internlm/Lean-Workbook.

☆ PALMS: Plane-based Accessible Indoor Localization Using Mobile Smartphones

Yunqian Cheng, Roberto Manduchi

In this paper, we present PALMS, an innovative indoor global localization and relocalization system for mobile smartphones that utilizes publicly available floor plans. Unlike most vision-based methods that require constant visual input, our system adopts a dynamic form of localization that considers a single instantaneous observation and odometry data. The core contribution of this work is the introduction of a particle filter initialization method that leverages the Certainly Empty Space (CES) constraint along with principal orientation matching. This approach creates a spatial probability distribution of the device's location, significantly improving localization accuracy and reducing particle filter convergence time. Our experimental evaluations demonstrate that PALMS outperforms traditional methods with uniformly initialized particle filters, providing a more efficient and accessible approach to indoor wayfinding. By eliminating the need for prior environmental fingerprinting, PALMS provides a scalable and practical approach to indoor navigation.

comment: 7 pages, 3 figures, accepted to the 14th International Conference on Indoor Positioning and Indoor Navigation (IPIN) 2024, Best Presentation Award

☆ Geographical Node Clustering and Grouping to Guarantee Data IIDness in Federated Learning

Minkwon Lee, Hyoil Kim, Changhee Joo

Federated learning (FL) is a decentralized AI mechanism suitable for a large number of devices like in smart IoT. A major challenge of FL is the non-IID dataset problem, originating from the heterogeneous data collected by FL participants, leading to performance deterioration of the trained global model. There have been various attempts to rectify non-IID dataset, mostly focusing on manipulating the collected data. This paper, however, proposes a novel approach to ensure data IIDness by properly clustering and grouping mobile IoT nodes exploiting their geographical characteristics, so that each FL group can achieve IID dataset. We first provide an experimental evidence for the independence and identicalness features of IoT data according to the inter-device distance, and then propose Dynamic Clustering and Partial-Steady Grouping algorithms that partition FL participants to achieve near-IIDness in their dataset while considering device mobility. Our mechanism significantly outperforms benchmark grouping algorithms at least by 110 times in terms of the joint cost between the number of dropout devices and the evenness in per-group device count, with a mild increase in the number of groups only by up to 0.93 groups.

comment: 10 pages, 7 figures

☆ NetSafe: Exploring the Topological Safety of Multi-agent Networks

Miao Yu, Shilong Wang, Guibin Zhang, Junyuan Mao, Chenlong Yin, Qijiong Liu, Qingsong Wen, Kun Wang, Yang Wang

Large language models (LLMs) have empowered nodes within multi-agent networks with intelligence, showing growing applications in both academia and industry. However, how to prevent these networks from generating malicious information remains unexplored with previous research on single LLM's safety be challenging to transfer. In this paper, we focus on the safety of multi-agent networks from a topological perspective, investigating which topological properties contribute to safer networks. To this end, we propose a general framework, NetSafe along with an iterative RelCom interaction to unify existing diverse LLM-based agent frameworks, laying the foundation for generalized topological safety research. We identify several critical phenomena when multi-agent networks are exposed to attacks involving misinformation, bias, and harmful information, termed as Agent Hallucination and Aggregation Safety. Furthermore, we find that highly connected networks are more susceptible to the spread of adversarial attacks, with task performance in a Star Graph Topology decreasing by 29.7%. Besides, our proposed static metrics aligned more closely with real-world dynamic evaluations than traditional graph-theoretic metrics, indicating that networks with greater average distances from attackers exhibit enhanced safety. In conclusion, our work introduces a new topological perspective on the safety of LLM-based multi-agent networks and discovers several unreported phenomena, paving the way for future research to explore the safety of such networks.

☆ Revealing and Mitigating the Local Pattern Shortcuts of Mamba

Wangjie You, Zecheng Tang, Juntao Li, Lili Yao, Min Zhang

Large language models (LLMs) have advanced significantly due to the attention mechanism, but their quadratic complexity and linear memory demands limit their performance on long-context tasks. Recently, researchers introduced Mamba, an advanced model built upon State Space Models(SSMs) that offers linear complexity and constant memory. Although Mamba is reported to match or surpass the performance of attention-based models, our analysis reveals a performance gap: Mamba excels in tasks that involve localized key information but faces challenges with tasks that require handling distributed key information. Our controlled experiments suggest that this inconsistency arises from Mamba's reliance on local pattern shortcuts, which enable the model to remember local key information within its limited memory but hinder its ability to retain more dispersed information. Therefore, we introduce a global selection module into the Mamba model to address this issue. Experiments on both existing and proposed synthetic tasks, as well as real-world tasks, demonstrate the effectiveness of our method. Notably, with the introduction of only 4M extra parameters, our approach enables the Mamba model(130M) to achieve a significant improvement on tasks with distributed information, increasing its performance from 0 to 80.54 points.

☆ Learning to Generate and Evaluate Fact-checking Explanations with Transformers

Darius Feher, Abdullah Khered, Hao Zhang, Riza Batista-Navarro, Viktor Schlegel

In an era increasingly dominated by digital platforms, the spread of misinformation poses a significant challenge, highlighting the need for solutions capable of assessing information veracity. Our research contributes to the field of Explainable Artificial Antelligence (XAI) by developing transformer-based fact-checking models that contextualise and justify their decisions by generating human-accessible explanations. Importantly, we also develop models for automatic evaluation of explanations for fact-checking verdicts across different dimensions such as \texttt{(self)-contradiction}, \texttt{hallucination}, \texttt{convincingness} and \texttt{overall quality}. By introducing human-centred evaluation methods and developing specialised datasets, we emphasise the need for aligning Artificial Intelligence (AI)-generated explanations with human judgements. This approach not only advances theoretical knowledge in XAI but also holds practical implications by enhancing the transparency, reliability and users' trust in AI-driven fact-checking systems. Furthermore, the development of our metric learning models is a first step towards potentially increasing efficiency and reducing reliance on extensive manual assessment. Based on experimental results, our best performing generative model \textsc{ROUGE-1} score of 47.77, demonstrating superior performance in generating fact-checking explanations, particularly when provided with high-quality evidence. Additionally, the best performing metric learning model showed a moderately strong correlation with human judgements on objective dimensions such as \texttt{(self)-contradiction and \texttt{hallucination}, achieving a Matthews Correlation Coefficient (MCC) of around 0.7.}

comment: Forthcoming in Engineering Applications of Artificial Intelligence

☆ RAC: Efficient LLM Factuality Correction with Retrieval Augmentation

Changmao Li, Jeffrey Flanigan

Large Language Models (LLMs) exhibit impressive results across a wide range of natural language processing (NLP) tasks, yet they can often produce factually incorrect outputs. This paper introduces a simple but effective low-latency post-correction method, \textbf{Retrieval Augmented Correction (RAC)}, aimed at enhancing the factual performance of LLMs without requiring additional fine-tuning. Our method is general and can be used with any instruction-tuned LLM, and has greatly reduced latency compared to prior approaches. RAC decomposes the LLM's output into atomic facts and applies a fine-grained verification and correction process with retrieved content to verify and correct the LLM-generated output. Our extensive experiments show that RAC yields up to 30\% improvements over state-of-the-art baselines across two popular factuality evaluation datasets, validating its efficacy and robustness in both with and without the integration of Retrieval-Augmented Generation (RAG) across different LLMs.\footnote{Our code is at \url{https://github.com/jlab-nlp/Retrieval-Augmented-Correction}}

☆ Long Term Memory: The Foundation of AI Self-Evolution

Xun Jiang, Feng Li, Han Zhao, Jiaying Wang, Jun Shao, Shihao Xu, Shu Zhang, Weiling Chen, Xavier Tang, Yize Chen, Mengyue Wu, Weizhi Ma, Mengdi Wang, Tianqiao Chen

Large language models (LLMs) like GPTs, trained on vast datasets, have demonstrated impressive capabilities in language understanding, reasoning, and planning, achieving human-level performance in various tasks. Most studies focus on enhancing these models by training on ever-larger datasets to build more powerful foundation models. While training stronger models is important, enabling models to evolve during inference is equally crucial, a process we refer to as AI self-evolution. Unlike large-scale training, self-evolution may rely on limited data or interactions. Inspired by the columnar organization of the human cerebral cortex, we hypothesize that AI models could develop cognitive abilities and build internal representations through iterative interactions with their environment. To achieve this, models need long-term memory (LTM) to store and manage processed interaction data. LTM supports self-evolution by representing diverse experiences across environments and agents. In this report, we explore AI self-evolution and its potential to enhance models during inference. We examine LTM's role in lifelong learning, allowing models to evolve based on accumulated interactions. We outline the structure of LTM and the systems needed for effective data retention and representation. We also classify approaches for building personalized models with LTM data and show how these models achieve self-evolution through interaction. Using LTM, our multi-agent framework OMNE achieved first place on the GAIA benchmark, demonstrating LTM's potential for AI self-evolution. Finally, we present a roadmap for future research, emphasizing the importance of LTM for advancing AI technology and its practical applications.

comment: 56 pages, 13 figures

☆ LightFusionRec: Lightweight Transformers-Based Cross-Domain Recommendation Model

Vansh Kharidia, Dhruvi Paprunia, Prashasti Kanikar

This paper presents LightFusionRec, a novel lightweight cross-domain recommendation system that integrates DistilBERT for textual feature extraction and FastText for genre embedding. Important issues in recommendation systems, such as data sparsity, computational efficiency, and cold start issues, are addressed in methodology. LightFusionRec uses a small amount of information to produce precise and contextually relevant recommendations for many media formats by fusing genre vector embedding with natural language processing algorithms. Tests conducted on extensive movie and book datasets show notable enhancements in suggestion quality when compared to conventional methods. Because of its lightweight design, the model can be used for a variety of purposes and allows for ondevice inference. LightFusionRec is a noteworthy development in cross-domain recommendation systems, providing accurate and scalable recommendations to improve user experience on digital content platforms.

☆ Opportunities and Challenges of Generative-AI in Finance

Akshar Prabhu Desai, Ganesh Satish Mallya, Mohammad Luqman, Tejasvi Ravi, Nithya Kota, Pranjul Yadav

Machine Learning and data mining have created widespread impact across various domains. However, these techniques are limited in their ability to reason, understand and generalize w.r.t language specific tasks. The aforementioned challenges were overcome, with the advancement of LLMs/Gen-AI. Gen-AI techniques are able to improve understanding of context and nuances in language modeling, translation between languages, handle large volumes of data, provide fast, low-latency responses and can be fine-tuned for various tasks and domains. In this manuscript, we present a comprehensive overview of the applications of Gen-AI techniques in the finance domain. In particular, we present the opportunities and challenges associated with the usage of Gen-AI techniques in finance. We also illustrate the various methodologies which can be used to train Gen-AI and present the various application areas of Gen-AI techniques in the finance ecosystem. To the best of our knowledge, this work represents the most comprehensive summarization of Gen-AI techniques within the financial domain. The analysis is designed for a deep overview of areas marked for substantial advancement while simultaneously pin-point those warranting future prioritization. We also hope that this work would serve as a conduit between finance and other domains, thus fostering the cross-pollination of innovative concepts and practices.

☆ Voice-Enabled AI Agents can Perform Common Scams

Richard Fang, Dylan Bowman, Daniel Kang

Recent advances in multi-modal, highly capable LLMs have enabled voice-enabled AI agents. These agents are enabling new applications, such as voice-enabled autonomous customer service. However, with all AI capabilities, these new capabilities have the potential for dual use. In this work, we show that voice-enabled AI agents can perform the actions necessary to perform common scams. To do so, we select a list of common scams collected by the government and construct voice-enabled agents with directions to perform these scams. We conduct experiments on our voice-enabled agents and show that they can indeed perform the actions necessary to autonomously perform such scams. Our results raise questions around the widespread deployment of voice-enabled AI agents.

☆ Boosting Jailbreak Transferability for Large Language Models

Hanqing Liu, Lifeng Zhou, Huanqian Yan

Large language models have drawn significant attention to the challenge of safe alignment, especially regarding jailbreak attacks that circumvent security measures to produce harmful content. To address the limitations of existing methods like GCG, which perform well in single-model attacks but lack transferability, we propose several enhancements, including a scenario induction template, optimized suffix selection, and the integration of re-suffix attack mechanism to reduce inconsistent outputs. Our approach has shown superior performance in extensive experiments across various benchmarks, achieving nearly 100% success rates in both attack execution and transferability. Notably, our method has won the online first place in the AISG-hosted Global Challenge for Safe and Secure LLMs.

☆ Procedural Content Generation in Games: A Survey with Insights on Emerging LLM Integration

Mahdi Farrokhi Maleki, Richard Zhao

Procedural Content Generation (PCG) is defined as the automatic creation of game content using algorithms. PCG has a long history in both the game industry and the academic world. It can increase player engagement and ease the work of game designers. While recent advances in deep learning approaches in PCG have enabled researchers and practitioners to create more sophisticated content, it is the arrival of Large Language Models (LLMs) that truly disrupted the trajectory of PCG advancement. This survey explores the differences between various algorithms used for PCG, including search-based methods, machine learning-based methods, other frequently used methods (e.g., noise functions), and the newcomer, LLMs. We also provide a detailed discussion on combined methods. Furthermore, we compare these methods based on the type of content they generate and the publication dates of their respective papers. Finally, we identify gaps in the existing academic work and suggest possible directions for future research.

☆ Resource-Efficient Medical Report Generation using Large Language Models

Abdullah, Ameer Hamza, Seong Tae Kim

Medical report generation is the task of automatically writing radiology reports for chest X-ray images. Manually composing these reports is a time-consuming process that is also prone to human errors. Generating medical reports can therefore help reduce the burden on radiologists. In other words, we can promote greater clinical automation in the medical domain. In this work, we propose a new framework leveraging vision-enabled Large Language Models (LLM) for the task of medical report generation. We introduce a lightweight solution that achieves better or comparative performance as compared to previous solutions on the task of medical report generation. We conduct extensive experiments exploring different model sizes and enhancement approaches, such as prefix tuning to improve the text generation abilities of the LLMs. We evaluate our approach on a prominent large-scale radiology report dataset - MIMIC-CXR. Our results demonstrate the capability of our resource-efficient framework to generate patient-specific reports with strong medical contextual understanding and high precision.

☆ Selecting Influential Samples for Long Context Alignment via Homologous Models' Guidance and Contextual Awareness Measurement

Shuzheng Si, Haozhe Zhao, Gang Chen, Yunshui Li, Kangyang Luo, Chuancheng Lv, Kaikai An, Fanchao Qi, Baobao Chang, Maosong Sun

The expansion of large language models to effectively handle instructions with extremely long contexts has yet to be fully investigated. The primary obstacle lies in constructing a high-quality long instruction-following dataset devised for long context alignment. Existing studies have attempted to scale up the available data volume by synthesizing long instruction-following samples. However, indiscriminately increasing the quantity of data without a well-defined strategy for ensuring data quality may introduce low-quality samples and restrict the final performance. To bridge this gap, we aim to address the unique challenge of long-context alignment, i.e., modeling the long-range dependencies for handling instructions and lengthy input contexts. We propose GATEAU, a novel framework designed to identify the influential and high-quality samples enriched with long-range dependency relations by utilizing crafted Homologous Models' Guidance (HMG) and Contextual Awareness Measurement (CAM). Specifically, HMG attempts to measure the difficulty of generating corresponding responses due to the long-range dependencies, using the perplexity scores of the response from two homologous models with different context windows. Also, the role of CAM is to measure the difficulty of understanding the long input contexts due to long-range dependencies by evaluating whether the model's attention is focused on important segments. Built upon both proposed methods, we select the most challenging samples as the influential data to effectively frame the long-range dependencies, thereby achieving better performance of LLMs. Comprehensive experiments indicate that GATEAU effectively identifies samples enriched with long-range dependency relations and the model trained on these selected samples exhibits better instruction-following and long-context understanding capabilities.

☆ Towards Kriging-informed Conditional Diffusion for Regional Sea-Level Data Downscaling

Subhankar Ghosh, Arun Sharma, Jayant Gupta, Aneesh Subramanian, Shashi Shekhar

Given coarser-resolution projections from global climate models or satellite data, the downscaling problem aims to estimate finer-resolution regional climate data, capturing fine-scale spatial patterns and variability. Downscaling is any method to derive high-resolution data from low-resolution variables, often to provide more detailed and local predictions and analyses. This problem is societally crucial for effective adaptation, mitigation, and resilience against significant risks from climate change. The challenge arises from spatial heterogeneity and the need to recover finer-scale features while ensuring model generalization. Most downscaling methods \cite{Li2020} fail to capture the spatial dependencies at finer scales and underperform on real-world climate datasets, such as sea-level rise. We propose a novel Kriging-informed Conditional Diffusion Probabilistic Model (Ki-CDPM) to capture spatial variability while preserving fine-scale features. Experimental results on climate data show that our proposed method is more accurate than state-of-the-art downscaling techniques.

☆ Improving Parallel Program Performance Through DSL-Driven Code Generation with LLM Optimizers

Anjiang Wei, Allen Nie, Thiago S. F. X. Teixeira, Rohan Yadav, Wonchan Lee, Ke Wang, Alex Aiken

Mapping computations to processors and assigning data to memory are critical for maximizing performance in parallel programming. These mapping decisions are managed through the development of specialized low-level system code, called mappers, crafted by performance engineers. Each mapper is tailored to a specific application and optimized for the underlying machine architecture, a process that requires days of refinement and tuning from an expert. Despite advances in system research, automating mapper generation remains a challenge due to the complexity of making millions of decisions to find the optimal solution and generate the solution as code. We introduce an approach that leverages recent advances in LLM-based optimizers for mapper design. In under ten minutes, our method automatically discovers mappers that surpass human expert designs in scientific applications by up to 1.34X speedup. For parallel matrix multiplication algorithms, our mapper achieves up to 1.31X of the expert-designed solution. To achieve this, we simplify the complexity of low-level code generation by introducing a domain-specific language (DSL) that abstracts the low-level system programming details and defines a structured search space for LLMs to explore. To maximize the application performance, we use an LLM optimizer to improve an agentic system that generates the mapper code. As a result, this approach significantly reduces the workload for performance engineers while achieving substantial performance gains across diverse applications. Finally, our results demonstrate the effectiveness of LLM-based optimization in system design and suggest its potential for addressing other complex system challenges.

comment: 26 pages, 8 figures

☆ Weighted Diversified Sampling for Efficient Data-Driven Single-Cell Gene-Gene Interaction Discovery

Yifan Wu, Yuntao Yang, Zirui Liu, Zhao Li, Khushbu Pahwa, Rongbin Li, Wenjin Zheng, Xia Hu, Zhaozhuo Xu

Gene-gene interactions play a crucial role in the manifestation of complex human diseases. Uncovering significant gene-gene interactions is a challenging task. Here, we present an innovative approach utilizing data-driven computational tools, leveraging an advanced Transformer model, to unearth noteworthy gene-gene interactions. Despite the efficacy of Transformer models, their parameter intensity presents a bottleneck in data ingestion, hindering data efficiency. To mitigate this, we introduce a novel weighted diversified sampling algorithm. This algorithm computes the diversity score of each data sample in just two passes of the dataset, facilitating efficient subset generation for interaction discovery. Our extensive experimentation demonstrates that by sampling a mere 1\% of the single-cell dataset, we achieve performance comparable to that of utilizing the entire dataset.

☆ Reinforced Imitative Trajectory Planning for Urban Automated Driving

Di Zeng, Ling Zheng, Xiantong Yang, Yinong Li

Reinforcement learning (RL) faces challenges in trajectory planning for urban automated driving due to the poor convergence of RL and the difficulty in designing reward functions. The convergence problem is alleviated by combining RL with supervised learning. However, most existing approaches only reason one step ahead and lack the capability to plan for multiple future steps. Besides, although inverse reinforcement learning holds promise for solving the reward function design issue, existing methods for automated driving impose a linear structure assumption on reward functions, making them difficult to apply to urban automated driving. In light of these challenges, this paper proposes a novel RL-based trajectory planning method that integrates RL with imitation learning to enable multi-step planning. Furthermore, a transformer-based Bayesian reward function is developed, providing effective reward signals for RL in urban scenarios. Moreover, a hybrid-driven trajectory planning framework is proposed to enhance safety and interpretability. The proposed methods were validated on the large-scale real-world urban automated driving nuPlan dataset. The results demonstrated the significant superiority of the proposed methods over the baselines in terms of the closed-loop metrics. The code is available at https://github.com/Zigned/nuplan_zigned.

comment: 19 pages, 9 figures

☆ Deep Active Learning with Manifold-preserving Trajectory Sampling

Yingrui Ji, Vijaya Sindhoori Kaza, Nishanth Artham, Tianyang Wang

Active learning (AL) is for optimizing the selection of unlabeled data for annotation (labeling), aiming to enhance model performance while minimizing labeling effort. The key question in AL is which unlabeled data should be selected for annotation. Existing deep AL methods arguably suffer from bias incurred by clabeled data, which takes a much lower percentage than unlabeled data in AL context. We observe that such an issue is severe in different types of data, such as vision and non-vision data. To address this issue, we propose a novel method, namely Manifold-Preserving Trajectory Sampling (MPTS), aiming to enforce the feature space learned from labeled data to represent a more accurate manifold. By doing so, we expect to effectively correct the bias incurred by labeled data, which can cause a biased selection of unlabeled data. Despite its focus on manifold, the proposed method can be conveniently implemented by performing distribution mapping with MMD (Maximum Mean Discrepancies). Extensive experiments on various vision and non-vision benchmark datasets demonstrate the superiority of our method. Our source code can be found here.

☆ P-YOLOv8: Efficient and Accurate Real-Time Detection of Distracted Driving

Mohamed R. Elshamy, Heba M. Emara, Mohamed R. Shoaib, Abdel-Hameed A. Badawy

Distracted driving is a critical safety issue that leads to numerous fatalities and injuries worldwide. This study addresses the urgent need for efficient and real-time machine learning models to detect distracted driving behaviors. Leveraging the Pretrained YOLOv8 (P-YOLOv8) model, a real-time object detection system is introduced, optimized for both speed and accuracy. This approach addresses the computational constraints and latency limitations commonly associated with conventional detection models. The study demonstrates P-YOLOv8 versatility in both object detection and image classification tasks using the Distracted Driver Detection dataset from State Farm, which includes 22,424 images across ten behavior categories. Our research explores the application of P-YOLOv8 for image classification, evaluating its performance compared to deep learning models such as VGG16, VGG19, and ResNet. Some traditional models often struggle with low accuracy, while others achieve high accuracy but come with high computational costs and slow detection speeds, making them unsuitable for real-time applications. P-YOLOv8 addresses these issues by achieving competitive accuracy with significant computational cost and efficiency advantages. In particular, P-YOLOv8 generates a lightweight model with a size of only 2.84 MB and a lower number of parameters, totaling 1,451,098, due to its innovative architecture. It achieves a high accuracy of 99.46 percent with this small model size, opening new directions for deployment on inexpensive and small embedded devices using Tiny Machine Learning (TinyML). The experimental results show robust performance, making P-YOLOv8 a cost-effective solution for real-time deployment. This study provides a detailed analysis of P-YOLOv8's architecture, training, and performance benchmarks, highlighting its potential for real-time use in detecting distracted driving.

☆ Patrol Security Game: Defending Against Adversary with Freedom in Attack Timing, Location, and Duration

Hao-Tsung Yang, Ting-Kai Weng, Ting-Yu Chang, Kin Sum Liu, Shan Lin, Jie Gao, Shih-Yu Tsai

We explored the Patrol Security Game (PSG), a robotic patrolling problem modeled as an extensive-form Stackelberg game, where the attacker determines the timing, location, and duration of their attack. Our objective is to devise a patrolling schedule with an infinite time horizon that minimizes the attacker's payoff. We demonstrated that PSG can be transformed into a combinatorial minimax problem with a closed-form objective function. By constraining the defender's strategy to a time-homogeneous first-order Markov chain (i.e., the patroller's next move depends solely on their current location), we proved that the optimal solution in cases of zero penalty involves either minimizing the expected hitting time or return time, depending on the attacker model, and that these solutions can be computed efficiently. Additionally, we observed that increasing the randomness in the patrol schedule reduces the attacker's expected payoff in high-penalty cases. However, the minimax problem becomes non-convex in other scenarios. To address this, we formulated a bi-criteria optimization problem incorporating two objectives: expected maximum reward and entropy. We proposed three graph-based algorithms and one deep reinforcement learning model, designed to efficiently balance the trade-off between these two objectives. Notably, the third algorithm can identify the optimal deterministic patrol schedule, though its runtime grows exponentially with the number of patrol spots. Experimental results validate the effectiveness and scalability of our solutions, demonstrating that our approaches outperform state-of-the-art baselines on both synthetic and real-world crime datasets.

comment: Under review of TCPS

☆ A Comprehensive Comparative Study of Individual ML Models and Ensemble Strategies for Network Intrusion Detection Systems

Ismail Bibers, Osvaldo Arreche, Mustafa Abdallah

The escalating frequency of intrusions in networked systems has spurred the exploration of new research avenues in devising artificial intelligence (AI) techniques for intrusion detection systems (IDS). Various AI techniques have been used to automate network intrusion detection tasks, yet each model possesses distinct strengths and weaknesses. Selecting the optimal model for a given dataset can pose a challenge, necessitating the exploration of ensemble methods to enhance generalization and applicability in network intrusion detection. This paper addresses this gap by conducting a comprehensive evaluation of diverse individual models and both simple and advanced ensemble methods for network IDS. We introduce an ensemble learning framework tailored for assessing individual models and ensemble methods in network intrusion detection tasks. Our framework encompasses the loading of input datasets, training of individual models and ensemble methods, and the generation of evaluation metrics. Furthermore, we incorporate all features across individual models and ensemble techniques. The study presents results for our framework, encompassing 14 methods, including various bagging, stacking, blending, and boosting techniques applied to multiple base learners such as decision trees, neural networks, and among others. We evaluate the framework using two distinct network intrusion datasets, RoEduNet-SIMARGL2021 and CICIDS-2017, each possessing unique characteristics. Additionally, we categorize AI models based on their performances on our evaluation metrics and via their confusion matrices. Our assessment demonstrates the efficacy of learning across most setups explored in this study. Furthermore, we contribute to the community by releasing our source codes, providing a foundational ensemble learning framework for network intrusion detection.

☆ A Comprehensive Survey of Datasets, Theories, Variants, and Applications in Direct Preference Optimization

Wenyi Xiao, Zechuan Wang, Leilei Gan, Shuai Zhao, Wanggui He, Luu Anh Tuan, Long Chen, Hao Jiang, Zhou Zhao, Fei Wu

With the rapid advancement of large language models (LLMs), aligning policy models with human preferences has become increasingly critical. Direct Preference Optimization (DPO) has emerged as a promising approach for alignment, acting as an RL-free alternative to Reinforcement Learning from Human Feedback (RLHF). Despite DPO's various advancements and inherent limitations, an in-depth review of these aspects is currently lacking in the literature. In this work, we present a comprehensive review of the challenges and opportunities in DPO, covering theoretical analyses, variants, relevant preference datasets, and applications. Specifically, we categorize recent studies on DPO based on key research questions to provide a thorough understanding of DPO's current landscape. Additionally, we propose several future research directions to offer insights on model alignment for the research community.

☆ AMPLE: Emotion-Aware Multimodal Fusion Prompt Learning for Fake News Detection

Xiaoman Xu, Xiangrun Li, Taihang Wang, Ye Jiang

Detecting fake news in large datasets is challenging due to its diversity and complexity, with traditional approaches often focusing on textual features while underutilizing semantic and emotional elements. Current methods also rely heavily on large annotated datasets, limiting their effectiveness in more nuanced analysis. To address these challenges, this paper introduces Emotion-\textbf{A}ware \textbf{M}ultimodal Fusion \textbf{P}rompt \textbf{L}\textbf{E}arning (\textbf{AMPLE}) framework to address the above issue by combining text sentiment analysis with multimodal data and hybrid prompt templates. This framework extracts emotional elements from texts by leveraging sentiment analysis tools. It then employs Multi-Head Cross-Attention (MCA) mechanisms and similarity-aware fusion methods to integrate multimodal data. The proposed AMPLE framework demonstrates strong performance on two public datasets in both few-shot and data-rich settings, with results indicating the potential of emotional aspects in fake news detection. Furthermore, the study explores the impact of integrating large language models with this method for text sentiment extraction, revealing substantial room for further improvement. The code can be found at :\url{https://github.com/xxm1215/MMM2025_few-shot/

☆ OpenMU: Your Swiss Army Knife for Music Understanding

Mengjie Zhao, Zhi Zhong, Zhuoyuan Mao, Shiqi Yang, Wei-Hsiang Liao, Shusuke Takahashi, Hiromi Wakaki, Yuki Mitsufuji

We present OpenMU-Bench, a large-scale benchmark suite for addressing the data scarcity issue in training multimodal language models to understand music. To construct OpenMU-Bench, we leveraged existing datasets and bootstrapped new annotations. OpenMU-Bench also broadens the scope of music understanding by including lyrics understanding and music tool usage. Using OpenMU-Bench, we trained our music understanding model, OpenMU, with extensive ablations, demonstrating that OpenMU outperforms baseline models such as MU-Llama. Both OpenMU and OpenMU-Bench are open-sourced to facilitate future research in music understanding and to enhance creative music production efficiency.

comment: Resources: https://github.com/mzhaojp22/openmu

☆ Leveraging Retrieval-Augmented Generation for Culturally Inclusive Hakka Chatbots: Design Insights and User Perceptions RAS

Chen-Chi Chang, Han-Pi Chang, Hung-Shin Lee

In an era where cultural preservation is increasingly intertwined with technological innovation, this study introduces a groundbreaking approach to promoting and safeguarding the rich heritage of Taiwanese Hakka culture through the development of a Retrieval-Augmented Generation (RAG)-enhanced chatbot. Traditional large language models (LLMs), while powerful, often fall short in delivering accurate and contextually rich responses, particularly in culturally specific domains. By integrating external databases with generative AI models, RAG technology bridges this gap, empowering chatbots to not only provide precise answers but also resonate deeply with the cultural nuances that are crucial for authentic interactions. This study delves into the intricate process of augmenting the chatbot's knowledge base with targeted cultural data, specifically curated to reflect the unique aspects of Hakka traditions, language, and practices. Through dynamic information retrieval, the RAG-enhanced chatbot becomes a versatile tool capable of handling complex inquiries that demand an in-depth understanding of Hakka cultural context. This is particularly significant in an age where digital platforms often dilute cultural identities, making the role of culturally aware AI systems more critical than ever. System usability studies conducted as part of our research reveal a marked improvement in both user satisfaction and engagement, highlighting the chatbot's effectiveness in fostering a deeper connection with Hakka culture. The feedback underscores the potential of RAG technology to not only enhance user experience but also to serve as a vital instrument in the broader mission of ethnic mainstreaming and cultural celebration.

comment: Accepted to IEEE RASSE 2024

☆ Stacking Small Language Models for Generalizability

Laurence Liang

Recent advances show that large language models (LLMs) generalize strong performance across different natural language benchmarks. However, the large size of LLMs makes training and inference expensive and impractical to run in resource-limited settings. This paper introduces a new approach called fine-tuning stacks of language models (FSLM), which involves stacking small language models (SLM) as an alternative to LLMs. By fine-tuning each SLM to perform a specific task, this approach breaks down high level reasoning into multiple lower-level steps that specific SLMs are responsible for. As a result, FSLM allows for lower training and inference costs, and also improves model interpretability as each SLM communicates with the subsequent one through natural language. By evaluating FSLM on common natural language benchmarks, this paper highlights promising early results toward generalizable performance using FSLM as a cost-effective alternative to LLMs.

☆ Pruning Foundation Models for High Accuracy without Retraining EMNLP 2024

Pu Zhao, Fei Sun, Xuan Shen, Pinrui Yu, Zhenglun Kong, Yanzhi Wang, Xue Lin

Despite the superior performance, it is challenging to deploy foundation models or large language models (LLMs) due to their massive parameters and computations. While pruning is a promising technique to reduce model size and accelerate the inference, the traditional pruning techniques can hardly be applied for LLMs as they need to finetune the model on the full dataset with multiple epochs consuming massive data and hardware resources. To deal with this problem, post-training pruning methods are proposed to prune LLMs in one-shot without retraining. However, their accuracy after pruning may suffer from certain performance degradation due to the lack of retraining with massive data. To address this issue, in this paper, we first formulate the post-training problem for layer-wise LLM compression to simultaneously prune multiple weights in LLMs. Next, we provide an optimal solution for this problem and design our post-training pruning algorithm for both unstructured and semi-structured sparsity. Our extensive experiments demonstrate the superior performance of the proposed methods in comparison to SOTA baselines across various LLM families including transformer-based LLMs and Mamba-based LLMs. Code link: https://github.com/piuzha/APT

comment: Accepted by EMNLP 2024 findings

☆ Bayesian Concept Bottleneck Models with LLM Priors

Jean Feng, Avni Kothari, Luke Zier, Chandan Singh, Yan Shuo Tan

Concept Bottleneck Models (CBMs) have been proposed as a compromise between white-box and black-box models, aiming to achieve interpretability without sacrificing accuracy. The standard training procedure for CBMs is to predefine a candidate set of human-interpretable concepts, extract their values from the training data, and identify a sparse subset as inputs to a transparent prediction model. However, such approaches are often hampered by the tradeoff between enumerating a sufficiently large set of concepts to include those that are truly relevant versus controlling the cost of obtaining concept extractions. This work investigates a novel approach that sidesteps these challenges: BC-LLM iteratively searches over a potentially infinite set of concepts within a Bayesian framework, in which Large Language Models (LLMs) serve as both a concept extraction mechanism and prior. BC-LLM is broadly applicable and multi-modal. Despite imperfections in LLMs, we prove that BC-LLM can provide rigorous statistical inference and uncertainty quantification. In experiments, it outperforms comparator methods including black-box models, converges more rapidly towards relevant concepts and away from spuriously correlated ones, and is more robust to out-of-distribution samples.

☆ A Plug-and-Play Fully On-the-Job Real-Time Reinforcement Learning Algorithm for a Direct-Drive Tandem-Wing Experiment Platforms Under Multiple Random Operating Conditions

Zhang Minghao, Song Bifeng, Yang Xiaojun, Wang Liang

The nonlinear and unstable aerodynamic interference generated by the tandem wings of such biomimetic systems poses substantial challenges for motion control, especially under multiple random operating conditions. To address these challenges, the Concerto Reinforcement Learning Extension (CRL2E) algorithm has been developed. This plug-and-play, fully on-the-job, real-time reinforcement learning algorithm incorporates a novel Physics-Inspired Rule-Based Policy Composer Strategy with a Perturbation Module alongside a lightweight network optimized for real-time control. To validate the performance and the rationality of the module design, experiments were conducted under six challenging operating conditions, comparing seven different algorithms. The results demonstrate that the CRL2E algorithm achieves safe and stable training within the first 500 steps, improving tracking accuracy by 14 to 66 times compared to the Soft Actor-Critic, Proximal Policy Optimization, and Twin Delayed Deep Deterministic Policy Gradient algorithms. Additionally, CRL2E significantly enhances performance under various random operating conditions, with improvements in tracking accuracy ranging from 8.3% to 60.4% compared to the Concerto Reinforcement Learning (CRL) algorithm. The convergence speed of CRL2E is 36.11% to 57.64% faster than the CRL algorithm with only the Composer Perturbation and 43.52% to 65.85% faster than the CRL algorithm when both the Composer Perturbation and Time-Interleaved Capability Perturbation are introduced, especially in conditions where the standard CRL struggles to converge. Hardware tests indicate that the optimized lightweight network structure excels in weight loading and average inference time, meeting real-time control requirements.

comment: 63 pages, 32 figures

♻ ☆ Toward Generalizing Visual Brain Decoding to Unseen Subjects

Xiangtao Kong, Kexin Huang, Ping Li, Lei Zhang

Visual brain decoding aims to decode visual information from human brain activities. Despite the great progress, one critical limitation of current brain decoding research lies in the lack of generalization capability to unseen subjects. Prior works typically focus on decoding brain activity of individuals based on the observation that different subjects exhibit different brain activities, while it remains unclear whether brain decoding can be generalized to unseen subjects. This study aims to answer this question. We first consolidate an image-fMRI dataset consisting of stimulus-image and fMRI-response pairs, involving 177 subjects in the movie-viewing task of the Human Connectome Project (HCP). This dataset allows us to investigate the brain decoding performance with the increase of participants. We then present a learning paradigm that applies uniform processing across all subjects, instead of employing different network heads or tokenizers for individuals as in previous methods, which can accommodate a large number of subjects to explore the generalization capability across different subjects. A series of experiments are conducted and we have the following findings. First, the network exhibits clear generalization capabilities with the increase of training subjects. Second, the generalization capability is common to popular network architectures (MLP, CNN and Transformer). Third, the generalization performance is affected by the similarity between subjects. Our findings reveal the inherent similarities in brain activities across individuals. With the emerging of larger and more comprehensive datasets, it is possible to train a brain decoding foundation model in the future. Codes and models can be found at https://github.com/Xiangtaokong/TGBD.

♻ ☆ RILe: Reinforced Imitation Learning

Mert Albaba, Sammy Christen, Thomas Langarek, Christoph Gebhardt, Otmar Hilliges, Michael J. Black

Reinforcement Learning has achieved significant success in generating complex behavior but often requires extensive reward function engineering. Adversarial variants of Imitation Learning and Inverse Reinforcement Learning offer an alternative by learning policies from expert demonstrations via a discriminator. However, these methods struggle in complex tasks where randomly sampling expert-like behaviors is challenging. This limitation stems from their reliance on policy-agnostic discriminators, which provide insufficient guidance for agent improvement, especially as task complexity increases and expert behavior becomes more distinct. We introduce RILe (Reinforced Imitation Learning environment), a novel trainer-student system that learns a dynamic reward function based on the student's performance and alignment with expert demonstrations. In RILe, the student learns an action policy while the trainer, using reinforcement learning, continuously updates itself via the discriminator's feedback to optimize the alignment between the student and the expert. The trainer optimizes for long-term cumulative rewards from the discriminator, enabling it to provide nuanced feedback that accounts for the complexity of the task and the student's current capabilities. This approach allows for greater exploration of agent actions by providing graduated feedback rather than binary expert/non-expert classifications. By reducing dependence on policy-agnostic discriminators, RILe enables better performance in complex settings where traditional methods falter, outperforming existing methods by 2x in complex simulated robot-locomotion tasks.

♻ ☆ Bidirectional Decoding: Improving Action Chunking via Closed-Loop Resampling

Yuejiang Liu, Jubayer Ibn Hamid, Annie Xie, Yoonho Lee, Maximilian Du, Chelsea Finn

Predicting and executing a sequence of actions without intermediate replanning, known as action chunking, is increasingly used in robot learning from human demonstrations. Yet, its reported effects on the learned policy are inconsistent: some studies find it crucial for achieving strong results, while others observe decreased performance. In this paper, we first dissect how action chunking impacts the divergence between a learner and a demonstrator. We find that action chunking allows the learner to better capture the temporal dependencies in demonstrations but at the cost of reduced reactivity in stochastic environments. To address this tradeoff, we propose Bidirectional Decoding (BID), a test-time inference algorithm that bridges action chunking with closed-loop operations. BID samples multiple predictions at each time step and searches for the optimal one based on two criteria: (i) backward coherence, which favors samples that align with previous decisions; (ii) forward contrast, which seeks samples of high likelihood for future plans. By coupling decisions within and across action chunks, BID promotes consistency over time while maintaining reactivity to unexpected changes. Experimental results show that BID boosts the performance of two state-of-the-art generative policies across seven simulation benchmarks and two real-world tasks. Code and videos are available at https://bid-robot.github.io.

comment: Project website: https://bid-robot.github.io/

♻ ☆ Are Large Language Models Chameleons? An Attempt to Simulate Social Surveys

Mingmeng Geng, Sihong He, Roberto Trotta

Can large language models (LLMs) simulate social surveys? To answer this question, we conducted millions of simulations in which LLMs were asked to answer subjective questions. A comparison of different LLM responses with the European Social Survey (ESS) data suggests that the effect of prompts on bias and variability is fundamental, highlighting major cultural, age, and gender biases. We further discussed statistical methods for measuring the difference between LLM answers and survey data and proposed a novel measure inspired by Jaccard similarity, as LLM-generated responses are likely to have a smaller variance. Our experiments also reveal that it is important to analyze the robustness and variability of prompts before using LLMs to simulate social surveys, as their imitation abilities are approximate at best.

comment: 17 pages

♻ ☆ IoT-Based Preventive Mental Health Using Knowledge Graphs and Standards for Better Well-Being

Amelie Gyrard, Seyedali Mohammadi, Manas Gaur, Antonio Kung

Sustainable Development Goals (SDGs) give the UN a road map for development with Agenda 2030 as a target. SDG3 "Good Health and Well-Being" ensures healthy lives and promotes well-being for all ages. Digital technologies can support SDG3. Burnout and even depression could be reduced by encouraging better preventive health. Due to the lack of patient knowledge and focus to take care of their health, it is necessary to help patients before it is too late. New trends such as positive psychology and mindfulness are highly encouraged in the USA. Digital Twins (DTs) can help with the continuous monitoring of emotion using physiological signals (e.g., collected via wearables). DTs facilitate monitoring and provide constant health insight to improve quality of life and well-being with better personalization. Healthcare DTs challenges are standardizing data formats, communication protocols, and data exchange mechanisms. As an example, ISO has the ISO/IEC JTC 1/SC 41 Internet of Things (IoT) and DTs Working Group, with standards such as "ISO/IEC 21823-3:2021 IoT - Interoperability for IoT Systems - Part 3 Semantic interoperability", "ISO/IEC CD 30178 - IoT - Data format, value and coding". To achieve those data integration and knowledge challenges, we designed the Mental Health Knowledge Graph (ontology and dataset) to boost mental health. As an example, explicit knowledge is described such as chocolate contains magnesium which is recommended for depression. The Knowledge Graph (KG) acquires knowledge from ontology-based mental health projects classified within the LOV4IoT ontology catalog (Emotion, Depression, and Mental Health). Furthermore, the KG is mapped to standards when possible. Standards from ETSI SmartM2M can be used such as SAREF4EHAW to represent medical devices and sensors, but also ITU/WHO, ISO, W3C, NIST, and IEEE standards relevant to mental health can be considered.

comment: 20 pages, Book chapter, Smart Technologies for Achieving Good Health and Well-Being: Towards Sustainable Development Goal, Taylor & Francis

♻ ☆ Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey

Tiansheng Huang, Sihao Hu, Fatih Ilhan, Selim Furkan Tekin, Ling Liu

Recent research demonstrates that the nascent fine-tuning-as-a-service business model exposes serious safety concerns -- fine-tuning over a few harmful data uploaded by the users can compromise the safety alignment of the model. The attack, known as harmful fine-tuning, has raised a broad research interest among the community. However, as the attack is still new, \textbf{we observe from our miserable submission experience that there are general misunderstandings within the research community.} We in this paper aim to clear some common concerns for the attack setting, and formally establish the research problem. Specifically, we first present the threat model of the problem, and introduce the harmful fine-tuning attack and its variants. Then we systematically survey the existing literature on attacks/defenses/mechanical analysis of the problem. Finally, we outline future research directions that might contribute to the development of the field. Additionally, we present a list of questions of interest, which might be useful to refer to when reviewers in the peer review process question the realism of the experiment/attack/defense setting. A curated list of relevant papers is maintained and made accessible at: \url{https://github.com/git-disl/awesome_LLM-harmful-fine-tuning-papers}.

♻ ☆ Adaptive $Q$-Network: On-the-fly Target Selection for Deep Reinforcement Learning

Théo Vincent, Fabian Wahren, Jan Peters, Boris Belousov, Carlo D'Eramo

Deep Reinforcement Learning (RL) is well known for being highly sensitive to hyperparameters, requiring practitioners substantial efforts to optimize them for the problem at hand. This also limits the applicability of RL in real-world scenarios. In recent years, the field of automated Reinforcement Learning (AutoRL) has grown in popularity by trying to address this issue. However, these approaches typically hinge on additional samples to select well-performing hyperparameters, hindering sample-efficiency and practicality. Furthermore, most AutoRL methods are heavily based on already existing AutoML methods, which were originally developed neglecting the additional challenges inherent to RL due to its non-stationarities. In this work, we propose a new approach for AutoRL, called Adaptive $Q$-Network (AdaQN), that is tailored to RL to take into account the non-stationarity of the optimization procedure without requiring additional samples. AdaQN learns several $Q$-functions, each one trained with different hyperparameters, which are updated online using the $Q$-function with the smallest approximation error as a shared target. Our selection scheme simultaneously handles different hyperparameters while coping with the non-stationarity induced by the RL optimization procedure and being orthogonal to any critic-based RL algorithm. We demonstrate that AdaQN is theoretically sound and empirically validate it in MuJoCo control problems and Atari $2600$ games, showing benefits in sample-efficiency, overall performance, robustness to stochasticity and training stability.

comment: Preprint

♻ ☆ RACCooN: A Versatile Instructional Video Editing Framework with Auto-Generated Narratives

Jaehong Yoon, Shoubin Yu, Mohit Bansal

Recent video generative models primarily rely on carefully written text prompts for specific tasks, like inpainting or style editing. They require labor-intensive textual descriptions for input videos, hindering their flexibility to adapt personal/raw videos to user specifications. This paper proposes RACCooN, a versatile and user-friendly video-to-paragraph-to-video generative framework that supports multiple video editing capabilities such as removal, addition, and modification, through a unified pipeline. RACCooN consists of two principal stages: Video-to-Paragraph (V2P) and Paragraph-to-Video (P2V). In the V2P stage, we automatically describe video scenes in well-structured natural language, capturing both the holistic context and focused object details. Subsequently, in the P2V stage, users can optionally refine these descriptions to guide the video diffusion model, enabling various modifications to the input video, such as removing, changing subjects, and/or adding new objects. The proposed approach stands out from other methods through several significant contributions: (1) RACCooN suggests a multi-granular spatiotemporal pooling strategy to generate well-structured video descriptions, capturing both the broad context and object details without requiring complex human annotations, simplifying precise video content editing based on text for users. (2) Our video generative model incorporates auto-generated narratives or instructions to enhance the quality and accuracy of the generated content. (3) RACCooN also plans to imagine new objects in a given video, so users simply prompt the model to receive a detailed video editing plan for complex video editing. The proposed framework demonstrates impressive versatile capabilities in video-to-paragraph generation, video content editing, and can be incorporated into other SoTA video generative models for further enhancement.

comment: The first two authors contribute equally. Project Page: https://raccoon-mllm-gen.github.io/

♻ ☆ Exploring the Potential of Large Language Models for Heterophilic Graphs

Yuxia Wu, Shujie Li, Yuan Fang, Chuan Shi

Large language models (LLMs) have presented significant opportunities to enhance various machine learning applications, including graph neural networks (GNNs). By leveraging the vast open-world knowledge within LLMs, we can more effectively interpret and utilize textual data to better characterize heterophilic graphs, where neighboring nodes often have different labels. However, existing approaches for heterophilic graphs overlook the rich textual data associated with nodes, which could unlock deeper insights into their heterophilic contexts. In this work, we explore the potential of LLMs for modeling heterophilic graphs and propose a novel two-stage framework: LLM-enhanced edge discriminator and LLM-guided edge reweighting. In the first stage, we fine-tune the LLM to better identify homophilic and heterophilic edges based on the textual content of their nodes. In the second stage, we adaptively manage message propagation in GNNs for different edge types based on node features, structures, and heterophilic or homophilic characteristics. To cope with the computational demands when deploying LLMs in practical scenarios, we further explore model distillation techniques to fine-tune smaller, more efficient models that maintain competitive performance. Extensive experiments validate the effectiveness of our framework, demonstrating the feasibility of using LLMs to enhance node classification on heterophilic graphs.

comment: Under review

♻ ☆ Do Large Language Models Need a Content Delivery Network?

Yihua Cheng, Kuntai Du, Jiayi Yao, Junchen Jiang

As the use of large language models (LLMs) expands rapidly, so does the range of knowledge needed to supplement various LLM queries. Thus, enabling flexible and efficient injection of new knowledge in LLM inference is critical. Three high-level options exist: (i) embedding the knowledge in LLM's weights (i.e., fine-tuning), (ii) including the knowledge as a part of LLM's text input (i.e., in-context learning), or (iii) injecting the KV caches of the new knowledge to LLM during prefill. This paper argues that, although fine-tuning and in-context learning are popular, using KV caches as the medium of knowledge could simultaneously enable more modular management of knowledge injection and more efficient LLM serving with low cost and fast response. To realize these benefits, we envision a Knowledge Delivery Network (KDN), a new system component in LLM services that dynamically optimizes the storage, transfer, and composition of KV cache across LLM engines and other compute and storage resources. We believe that, just like content delivery networks (CDNs), such as Akamai, enabled the success of the Internet ecosystem through their efficient data delivery, KDNs will be critical to the success of LLM applications through their efficient knowledge delivery. We have open-sourced a KDN prototype at https://github.com/LMCache/LMCache.

♻ ☆ MathGAP: Out-of-Distribution Evaluation on Problems with Arbitrarily Complex Proofs

Andreas Opedal, Haruki Shirakami, Bernhard Schölkopf, Abulhair Saparov, Mrinmaya Sachan

Large language models (LLMs) can solve arithmetic word problems with high accuracy, but little is known about how well they generalize to problems that are more complex than the ones on which they have been trained. Empirical investigations of such questions are impeded by two major flaws of current evaluations: (i) much of the evaluation data is contaminated, in the sense that it has already been seen during training, and (ii) benchmark datasets do not capture how problem proofs may be arbitrarily complex in various ways. As a step towards addressing these issues, we present a framework for evaluating LLMs on problems with arbitrarily complex arithmetic proofs, called MathGAP. MathGAP generates problems that follow fixed proof specifications -- along with chain-of-thought reasoning annotations -- enabling systematic studies on generalization with respect to arithmetic proof complexity. We apply MathGAP to analyze how in-context learning interacts with generalization to problems that have more complex proofs. We find that among the models tested, most show a significant decrease in performance as proofs get deeper and wider. This effect is more pronounced in complex, nonlinear proof structures, which are challenging even for GPT-4o. Surprisingly, providing in-context examples from the same distribution as the test set is not always beneficial for performance. In particular, zero-shot prompting as well as demonstrating a diverse range of examples that are less complex than the test data sometimes yield similar or higher accuracies.

comment: Preprint

♻ ☆ Human-Agent Joint Learning for Efficient Robot Manipulation Skill Acquisition

Shengcheng Luo, Quanquan Peng, Jun Lv, Kaiwen Hong, Katherine Rose Driggs-Campbell, Cewu Lu, Yong-Lu Li

Employing a teleoperation system for gathering demonstrations offers the potential for more efficient learning of robot manipulation. However, teleoperating a robot arm equipped with a dexterous hand or gripper, via a teleoperation system presents inherent challenges due to the task's high dimensionality, complexity of motion, and differences between physiological structures. In this study, we introduce a novel system for joint learning between human operators and robots, that enables human operators to share control of a robot end-effector with a learned assistive agent, simplifies the data collection process, and facilitates simultaneous human demonstration collection and robot manipulation training. As data accumulates, the assistive agent gradually learns. Consequently, less human effort and attention are required, enhancing the efficiency of the data collection process. It also allows the human operator to adjust the control ratio to achieve a trade-off between manual and automated control. We conducted experiments in both simulated environments and physical real-world settings. Through user studies and quantitative evaluations, it is evident that the proposed system could enhance data collection efficiency and reduce the need for human adaptation while ensuring the collected data is of sufficient quality for downstream tasks. \textit{For more details, please refer to our webpage https://norweig1an.github.io/HAJL.github.io/.

comment: 8 pages, 6 figures

♻ ☆ Feature Mapping in Physics-Informed Neural Networks (PINNs)

Chengxi Zeng, Tilo Burghardt, Alberto M Gambaruto

In this paper, the training dynamics of PINNs with a feature mapping layer via the limiting Conjugate Kernel and Neural Tangent Kernel is investigated, shedding light on the convergence of PINNs; Although the commonly used Fourier-based feature mapping has achieved great success, we show its inadequacy in some physics scenarios. Via these two scopes, we propose conditionally positive definite Radial Basis Function as a better alternative. Lastly, we explore the feature mapping numerically in wide neural networks. Our empirical results reveal the efficacy of our method in diverse forward and inverse problem sets. Composing feature functions is found to be a practical way to address the expressivity and generalisability trade-off, viz., tuning the bandwidth of the kernels and the surjectivity of the feature mapping function. This simple technique can be implemented for coordinate inputs and benefits the broader PINNs research.

♻ ☆ Proceedings of The second international workshop on eXplainable AI for the Arts (XAIxArts)

Nick Bryan-Kinns, Corey Ford, Shuoyang Zheng, Helen Kennedy, Alan Chamberlain, Makayla Lewis, Drew Hemment, Zijin Li, Qiong Wu, Lanxi Xiao, Gus Xia, Jeba Rezwana, Michael Clemens, Gabriel Vigliensoni

This second international workshop on explainable AI for the Arts (XAIxArts) brought together a community of researchers in HCI, Interaction Design, AI, explainable AI (XAI), and digital arts to explore the role of XAI for the Arts. Workshop held at the 16th ACM Conference on Creativity and Cognition (C&C 2024), Chicago, USA.

comment: Proceedings of The second international workshop on eXplainable AI for the Arts (XAIxArts)

♻ ☆ TabSeq: A Framework for Deep Learning on Tabular Data via Sequential Ordering ICPR 2024

Al Zadid Sultan Bin Habib, Kesheng Wang, Mary-Anne Hartley, Gianfranco Doretto, Donald A. Adjeroh

Effective analysis of tabular data still poses a significant problem in deep learning, mainly because features in tabular datasets are often heterogeneous and have different levels of relevance. This work introduces TabSeq, a novel framework for the sequential ordering of features, addressing the vital necessity to optimize the learning process. Features are not always equally informative, and for certain deep learning models, their random arrangement can hinder the model's learning capacity. Finding the optimum sequence order for such features could improve the deep learning models' learning process. The novel feature ordering technique we provide in this work is based on clustering and incorporates both local ordering and global ordering. It is designed to be used with a multi-head attention mechanism in a denoising autoencoder network. Our framework uses clustering to align comparable features and improve data organization. Multi-head attention focuses on essential characteristics, whereas the denoising autoencoder highlights important aspects by rebuilding from distorted inputs. This method improves the capability to learn from tabular data while lowering redundancy. Our research, demonstrating improved performance through appropriate feature sequence rearrangement using raw antibody microarray and two other real-world biomedical datasets, validates the impact of feature ordering. These results demonstrate that feature ordering can be a viable approach to improved deep learning of tabular data.

comment: This paper has been accepted for presentation at the 27th International Conference on Pattern Recognition (ICPR 2024) in Kolkata, India

♻ ☆ Generalized Group Data Attribution

Dan Ley, Suraj Srinivas, Shichang Zhang, Gili Rusak, Himabindu Lakkaraju

Data Attribution (DA) methods quantify the influence of individual training data points on model outputs and have broad applications such as explainability, data selection, and noisy label identification. However, existing DA methods are often computationally intensive, limiting their applicability to large-scale machine learning models. To address this challenge, we introduce the Generalized Group Data Attribution (GGDA) framework, which computationally simplifies DA by attributing to groups of training points instead of individual ones. GGDA is a general framework that subsumes existing attribution methods and can be applied to new DA techniques as they emerge. It allows users to optimize the trade-off between efficiency and fidelity based on their needs. Our empirical results demonstrate that GGDA applied to popular DA methods such as Influence Functions, TracIn, and TRAK results in upto 10x-50x speedups over standard DA methods while gracefully trading off attribution fidelity. For downstream applications such as dataset pruning and noisy label identification, we demonstrate that GGDA significantly improves computational efficiency and maintains effectiveness, enabling practical applications in large-scale machine learning scenarios that were previously infeasible.

♻ ☆ Latent Skill Discovery for Chain-of-Thought Reasoning

Zifan Xu, Haozhu Wang, Dmitriy Bespalov, Xuan Wang, Peter Stone, Yanjun Qi

Chain-of-thought (CoT) prompting is a popular in-context learning (ICL) approach for large language models (LLMs), especially when tackling complex reasoning tasks. Traditional ICL approaches construct prompts using examples that contain questions similar to the input question. However, CoT prompting, which includes crucial intermediate reasoning steps (rationales) within its examples, necessitates selecting examples based on these rationales rather than the questions themselves. Existing methods require human experts or pre-trained LLMs to describe the skill, a high-level abstraction of rationales, to guide the selection. These methods, however, are often costly and difficult to scale. Instead, this paper introduces a new approach named Latent Reasoning Skills (LaRS) that employs unsupervised learning to create a latent space representation of rationales, with a latent variable called a reasoning skill. Concurrently, LaRS learns a reasoning policy to determine the required reasoning skill for a given question. Then the ICL examples are selected by aligning the reasoning skills between past examples and the question. This approach is theoretically grounded and compute-efficient, eliminating the need for auxiliary LLM inference or manual prompt design. Empirical results demonstrate that LaRS consistently outperforms SOTA skill-based selection methods, processing example banks four times faster, reducing LLM inferences during the selection stage by half, and showing greater robustness to sub-optimal example banks.

♻ ☆ From FDG to PSMA: A Hitchhiker's Guide to Multitracer, Multicenter Lesion Segmentation in PET/CT Imaging

Maximilian Rokuss, Balint Kovacs, Yannick Kirchhoff, Shuhan Xiao, Constantin Ulrich, Klaus H. Maier-Hein, Fabian Isensee

Automated lesion segmentation in PET/CT scans is crucial for improving clinical workflows and advancing cancer diagnostics. However, the task is challenging due to physiological variability, different tracers used in PET imaging, and diverse imaging protocols across medical centers. To address this, the autoPET series was created to challenge researchers to develop algorithms that generalize across diverse PET/CT environments. This paper presents our solution for the autoPET III challenge, targeting multitracer, multicenter generalization using the nnU-Net framework with the ResEncL architecture. Key techniques include misalignment data augmentation and multi-modal pretraining across CT, MR, and PET datasets to provide an initial anatomical understanding. We incorporate organ supervision as a multitask approach, enabling the model to distinguish between physiological uptake and tracer-specific patterns, which is particularly beneficial in cases where no lesions are present. Compared to the default nnU-Net, which achieved a Dice score of 57.61, or the larger ResEncL (65.31) our model significantly improved performance with a Dice score of 68.40, alongside a reduction in false positive (FPvol: 7.82) and false negative (FNvol: 10.35) volumes. These results underscore the effectiveness of combining advanced network design, augmentation, pretraining, and multitask learning for PET/CT lesion segmentation. After evaluation on the test set, our approach was awarded the first place in the model-centric category (Team LesionTracer). Code is publicly available at https://github.com/MIC-DKFZ/autopet-3-submission.

comment: Winning method of the autoPET III challenge (model-centric) - Team LesionTracer

♻ ☆ Dynamics of Moral Behavior in Heterogeneous Populations of Learning Agents AAAI

Elizaveta Tennant, Stephen Hailes, Mirco Musolesi

Growing concerns about safety and alignment of AI systems highlight the importance of embedding moral capabilities in artificial agents: a promising solution is the use of learning from experience, i.e., Reinforcement Learning. In multi-agent (social) environments, complex population-level phenomena may emerge from interactions between individual learning agents. Many of the existing studies rely on simulated social dilemma environments to study the interactions of independent learning agents; however, they tend to ignore the moral heterogeneity that is likely to be present in societies of agents in practice. For example, at different points in time a single learning agent may face opponents who are consequentialist (i.e., focused on maximizing outcomes over time), norm-based (i.e., conforming to specific norms), or virtue-based (i.e., considering a combination of different virtues). The extent to which agents' co-development may be impacted by such moral heterogeneity in populations is not well understood. In this paper, we present a study of the learning dynamics of morally heterogeneous populations interacting in a social dilemma setting. Using an Iterated Prisoner's Dilemma environment with a partner selection mechanism, we investigate the extent to which the prevalence of diverse moral agents in populations affects individual agents' learning behaviors and emergent population-level outcomes. We observe several types of non-trivial interactions between pro-social and anti-social agents, and find that certain types of moral agents are able to steer selfish agents towards more cooperative behavior.

comment: Presented at AIES 2024 (7th AAAI/ACM Conference on AI, Ethics, and Society - San Jose, CA, USA) https://ojs.aaai.org/index.php/AIES/article/view/31736

♻ ★ DISCO: Efficient Diffusion Solver for Large-Scale Combinatorial Optimization Problems

Kexiong Yu, Hang Zhao, Yuhang Huang, Renjiao Yi, Kai Xu, Chenyang Zhu

Combinatorial Optimization (CO) problems are fundamentally important in numerous real-world applications across diverse industries, characterized by entailing enormous solution space and demanding time-sensitive response. Despite recent advancements in neural solvers, their limited expressiveness struggles to capture the multi-modal nature of CO landscapes. While some research has shifted towards diffusion models, these models still sample solutions indiscriminately from the entire NP-complete solution space with time-consuming denoising processes, which limit their practicality for large problem scales. We propose DISCO, an efficient DIffusion Solver for large-scale Combinatorial Optimization problems that excels in both solution quality and inference speed. DISCO's efficacy is twofold: First, it enhances solution quality by constraining the sampling space to a more meaningful domain guided by solution residues, while preserving the multi-modal properties of the output distributions. Second, it accelerates the denoising process through an analytically solvable approach, enabling solution sampling with minimal reverse-time steps and significantly reducing inference time. DISCO delivers strong performance on large-scale Traveling Salesman Problems and challenging Maximal Independent Set benchmarks, with inference time up to 5.28 times faster than other diffusion alternatives. By incorporating a divide-and-conquer strategy, DISCO can well generalize to solve unseen-scale problem instances, even surpassing models specifically trained for those scales.

♻ ☆ Optimizing BioTac Simulation for Realistic Tactile Perception IJCNN

Wadhah Zai El Amri, Nicolás Navarro-Guerrero

Tactile sensing presents a promising opportunity for enhancing the interaction capabilities of today's robots. BioTac is a commonly used tactile sensor that enables robots to perceive and respond to physical tactile stimuli. However, the sensor's non-linearity poses challenges in simulating its behavior. In this paper, we first investigate a BioTac simulation that uses temperature, force, and contact point positions to predict the sensor outputs. We show that training with BioTac temperature readings does not yield accurate sensor output predictions during deployment. Consequently, we tested three alternative models, i.e., an XGBoost regressor, a neural network, and a transformer encoder. We train these models without temperature readings and provide a detailed investigation of the window size of the input vectors. We demonstrate that we achieve statistically significant improvements over the baseline network. Furthermore, our results reveal that the XGBoost regressor and transformer outperform traditional feed-forward neural networks in this task. We make all our code and results available online on https://github.com/wzaielamri/Optimizing_BioTac_Simulation.

comment: 12 pages (including appendix), Accepted at the International Joint Conference on Neural Network (IJCNN) 2024, Yokohama, Japan. \c{opyright} 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media... (We refer to IEEE Copyrights)

♻ ☆ Selection-p: Self-Supervised Task-Agnostic Prompt Compression for Faithfulness and Transferability EMNLP 2024

Tsz Ting Chung, Leyang Cui, Lemao Liu, Xinting Huang, Shuming Shi, Dit-Yan Yeung

Large Language Models (LLMs) have demonstrated impressive capabilities in a wide range of natural language processing tasks when leveraging in-context learning. To mitigate the additional computational and financial costs associated with in-context learning, several prompt compression methods have been proposed to compress the in-context learning prompts. Despite their success, these methods face challenges with transferability due to model-specific compression, or rely on external training data, such as GPT-4. In this paper, we investigate the ability of LLMs to develop a unified compression method that discretizes uninformative tokens, utilizing a self-supervised pre-training technique. By introducing a small number of parameters during the continual pre-training, the proposed Selection-p produces a probability for each input token, indicating whether to preserve or discard it. Experiments show Selection-p achieves state-of-the-art performance across numerous classification tasks, achieving compression rates of up to 10 times while experiencing only a marginal 0.8% decrease in performance. Moreover, it exhibits superior transferability to different models compared to prior work. Additionally, we further analyze how Selection-p helps maintain performance on in-context learning with long contexts.

comment: 14 pages, 5 figures, 10 tables, EMNLP 2024 Findings

♻ ☆ OAEI-LLM: A Benchmark Dataset for Understanding Large Language Model Hallucinations in Ontology Matching

Zhangcheng Qiang, Kerry Taylor, Weiqing Wang, Jing Jiang

Hallucinations of large language models (LLMs) commonly occur in domain-specific downstream tasks, with no exception in ontology matching (OM). The prevalence of using LLMs for OM raises the need for benchmarks to better understand LLM hallucinations. The OAEI-LLM dataset is an extended version of the Ontology Alignment Evaluation Initiative (OAEI) datasets that evaluate LLM-specific hallucinations in OM tasks. We outline the methodology used in dataset construction and schema extension, and provide examples of potential use cases.

comment: 5 pages, 1 figure, 1 table

♻ ☆ VeLoRA: Memory Efficient Training using Rank-1 Sub-Token Projections NeurIPS 2024

Roy Miles, Pradyumna Reddy, Ismail Elezi, Jiankang Deng

Large language models (LLMs) have recently emerged as powerful tools for tackling many language-processing tasks. Despite their success, training and fine-tuning these models is still far too computationally and memory intensive. In this paper, we identify and characterise the important components needed for effective model convergence using gradient descent. In doing so we find that the intermediate activations used to implement backpropagation can be excessively compressed without incurring any degradation in performance. This result leads us to a cheap and memory-efficient algorithm for both fine-tuning and pre-training LLMs. The proposed algorithm simply divides the tokens up into smaller sub-tokens before projecting them onto a fixed 1-dimensional subspace during the forward pass. These features are then coarsely reconstructed during the backward pass to implement the update rules. We confirm the effectiveness of our algorithm as being complimentary to many state-of-the-art PEFT methods on the VTAB-1k fine-tuning benchmark. Furthermore, we outperform QLoRA for fine-tuning LLaMA and show competitive performance against other memory-efficient pre-training methods on the large-scale C4 dataset.

comment: NeurIPS 2024. Code available at https://github.com/roymiles/VeLoRA

♻ ☆ Compiled Models, Built-In Exploits: Uncovering Pervasive Bit-Flip Attack Surfaces in DNN Executables NDSS 2025

Yanzuo Chen, Zhibo Liu, Yuanyuan Yuan, Sihang Hu, Tianxiang Li, Shuai Wang

Bit-flip attacks (BFAs) can manipulate deep neural networks (DNNs). For high-level DNN models running on deep learning (DL) frameworks like PyTorch, extensive BFAs have been used to flip bits in model weights and shown effective. Defenses have also been proposed to guard model weights. However, DNNs are increasingly compiled into DNN executables by DL compilers to leverage hardware primitives. These executables manifest distinct computation paradigms; existing research fails to accurately capture and expose the BFA surfaces on DNN executables. To this end, we launch the first systematic study of BFAs on DNN executables. Prior BFAs are limited to attacking model weights and assume a strong whitebox attacker with full knowledge of victim model weights, which is unrealistic as weights are often confidential. In contrast, we find that BFAs on DNN executables can achieve high effectiveness by exploiting the model structure (usually stored in the executable code), which only requires knowing the (often public) model structure. Importantly, such structure-based BFAs are pervasive, transferable, and more severe in DNN executables. They also slip past existing defenses. To demonstrate the new attack surfaces, we assume a weak and more realistic attacker with no knowledge of victim model weights. We design an automated tool to identify vulnerable bits in victim executables with high confidence (70% vs. baseline 2%). We show on DDR4 DRAM that only 1.4 flips on average are needed to fully downgrade the accuracy of victim models, including quantized ones which could require 23x more flips previously, to random guesses. We comprehensively evaluate 16 DNN executables, covering large-scale models trained on commonly-used datasets compiled by the two most popular DL compilers. Our finding calls for incorporating security mechanisms in future DNN compilation toolchains.

comment: Accepted by NDSS 2025

♻ ☆ Towards Next-Level Post-Training Quantization of Hyper-Scale Transformers NeurIPS 2024

Junhan Kim, Chungman Lee, Eulrang Cho, Kyungphil Park, Ho-young Kim, Joonyoung Kim, Yongkweon Jeon

With the increasing complexity of generative AI models, post-training quantization (PTQ) has emerged as a promising solution for deploying hyper-scale models on edge devices such as mobile and TVs. Existing PTQ schemes, however, consume considerable time and resources, which could be a bottleneck in real situations where frequent model updates and multiple hyperparameter tunings are required. As a cost-effective alternative, learning-free PTQ schemes have been proposed. However, the performance is somewhat limited because they cannot consider the inter-layer dependency within the attention module, which is a significant feature of Transformers. In this paper, we thus propose a novel PTQ algorithm that balances accuracy and efficiency. The key idea of the proposed algorithm called aespa is to perform quantization layer-wise for efficiency while targeting attention-wise reconstruction to consider the cross-layer dependency. Through extensive experiments on various language models and complexity analysis, we demonstrate that aespa is accurate and efficient in quantizing Transformer models.

comment: Accepted to NeurIPS 2024

♻ ☆ UADA3D: Unsupervised Adversarial Domain Adaptation for 3D Object Detection with Sparse LiDAR and Large Domain Gaps

Maciej K Wozniak, Mattias Hansson, Marko Thiel, Patric Jensfelt

In this study, we address a gap in existing unsupervised domain adaptation approaches on LiDAR-based 3D object detection, which have predominantly concentrated on adapting between established, high-density autonomous driving datasets. We focus on sparser point clouds, capturing scenarios from different perspectives: not just from vehicles on the road but also from mobile robots on sidewalks, which encounter significantly different environmental conditions and sensor configurations. We introduce Unsupervised Adversarial Domain Adaptation for 3D Object Detection (UADA3D). UADA3D does not depend on pre-trained source models or teacher-student architectures. Instead, it uses an adversarial approach to directly learn domain-invariant features. We demonstrate its efficacy in various adaptation scenarios, showing significant improvements in both self-driving car and mobile robot domains. Our code is open-source and will be available soon.

comment: Accepted for IEEE RA-L 2024

♻ ☆ TrafficGamer: Reliable and Flexible Traffic Simulation for Safety-Critical Scenarios with Game-Theoretic Oracles

Guanren Qiao, Guorui Quan, Jiawei Yu, Shujun Jia, Guiliang Liu

While modern Autonomous Vehicle (AV) systems can develop reliable driving policies under regular traffic conditions, they frequently struggle with safety-critical traffic scenarios. This difficulty primarily arises from the rarity of such scenarios in driving datasets and the complexities associated with predictive modeling among multiple vehicles. To support the testing and refinement of AV policies, simulating safety-critical traffic events is an essential challenge to be addressed. In this work, we introduce TrafficGamer, which facilitates game-theoretic traffic simulation by viewing common road driving as a multi-agent game. In evaluating the empirical performance across various real-world datasets, TrafficGamer ensures both fidelity and exploitability of the simulated scenarios, guaranteeing that they not only statically align with real-world traffic distribution but also efficiently capture equilibriums for representing safety-critical scenarios involving multiple agents. Additionally, the results demonstrate that TrafficGamer exhibits highly flexible simulation across various contexts. Specifically, we demonstrate that the generated scenarios can dynamically adapt to equilibriums of varying tightness by configuring risk-sensitive constraints during optimization. To the best of our knowledge, TrafficGamer is the first simulator capable of generating diverse traffic scenarios involving multiple agents. We have provided a demo webpage for the project at https://qiaoguanren.github.io/trafficgamer-demo/.

♻ ☆ Log Probabilities Are a Reliable Estimate of Semantic Plausibility in Base and Instruction-Tuned Language Models

Carina Kauf, Emmanuele Chersoni, Alessandro Lenci, Evelina Fedorenko, Anna A. Ivanova

Semantic plausibility (e.g. knowing that "the actor won the award" is more likely than "the actor won the battle") serves as an effective proxy for general world knowledge. Language models (LMs) capture vast amounts of world knowledge by learning distributional patterns in text, accessible via log probabilities (LogProbs) they assign to plausible vs. implausible outputs. The new generation of instruction-tuned LMs can now also provide explicit estimates of plausibility via prompting. Here, we evaluate the effectiveness of LogProbs and basic prompting to measure semantic plausibility, both in single-sentence minimal pairs (Experiment 1) and short context-dependent scenarios (Experiment 2). We find that (i) in both base and instruction-tuned LMs, LogProbs offers a more reliable measure of semantic plausibility than direct zero-shot prompting, which yields inconsistent and often poor results; (ii) instruction-tuning generally does not alter the sensitivity of LogProbs to semantic plausibility (although sometimes decreases it); (iii) across models, context mostly modulates LogProbs in expected ways, as measured by three novel metrics of context-sensitive plausibility and their match to explicit human plausibility judgments. We conclude that, even in the era of prompt-based evaluations, LogProbs constitute a useful metric of semantic plausibility, both in base and instruction-tuned LMs.

♻ ☆ StrucText-Eval: Evaluating Large Language Model's Reasoning Ability in Structure-Rich Text

Zhouhong Gu, Haoning Ye, Xingzhou Chen, Zeyang Zhou, Hongwei Feng, Yanghua Xiao

The effective utilization of structured data, integral to corporate data strategies, has been challenged by the rise of large language models (LLMs) capable of processing unstructured information. This shift prompts the question: can LLMs interpret structured data directly in its unstructured form? We propose an automatic evaluation data generation method for assessing LLMs' reasoning capabilities on structure-rich text to explore this. Our approach supports 8 structured languages and 29 tasks, generating data with adjustable complexity through controllable nesting and structural width. We introduce StrucText-Eval, a benchmark containing 5,800 pre-generated and annotated samples designed to evaluate how well LLMs understand and reason through structured text. StrucText-Eval is divided into two suites: a regular Test suite (3,712 samples) and a Test-Hard suite (2,088 samples), the latter emphasizing the gap between human and model performance on more complex tasks. Experimental results show that while open-source LLMs achieve a maximum accuracy of 74.9\% on the standard dataset, their performance drops significantly to 45.8\% on the harder dataset. In contrast, human participants reach an accuracy of 92.6\% on StrucText-Eval-Hard, highlighting LLMs' current limitations in handling intricate structural information. The benchmark and generation codes are open sourced in \url{https://github.com/MikeGu721/StrucText-Eval}

♻ ☆ Optimus-1: Hybrid Multimodal Memory Empowered Agents Excel in Long-Horizon Tasks NeurIPS 2024

Zaijing Li, Yuquan Xie, Rui Shao, Gongwei Chen, Dongmei Jiang, Liqiang Nie

Building a general-purpose agent is a long-standing vision in the field of artificial intelligence. Existing agents have made remarkable progress in many domains, yet they still struggle to complete long-horizon tasks in an open world. We attribute this to the lack of necessary world knowledge and multimodal experience that can guide agents through a variety of long-horizon tasks. In this paper, we propose a Hybrid Multimodal Memory module to address the above challenges. It 1) transforms knowledge into Hierarchical Directed Knowledge Graph that allows agents to explicitly represent and learn world knowledge, and 2) summarises historical information into Abstracted Multimodal Experience Pool that provide agents with rich references for in-context learning. On top of the Hybrid Multimodal Memory module, a multimodal agent, Optimus-1, is constructed with dedicated Knowledge-guided Planner and Experience-Driven Reflector, contributing to a better planning and reflection in the face of long-horizon tasks in Minecraft. Extensive experimental results show that Optimus-1 significantly outperforms all existing agents on challenging long-horizon task benchmarks, and exhibits near human-level performance on many tasks. In addition, we introduce various Multimodal Large Language Models (MLLMs) as the backbone of Optimus-1. Experimental results show that Optimus-1 exhibits strong generalization with the help of the Hybrid Multimodal Memory module, outperforming the GPT-4V baseline on many tasks.

comment: Accepted by NeurIPS 2024

♻ ☆ An Electrocardiogram Foundation Model Built on over 10 Million Recordings with External Evaluation across Multiple Domains

Jun Li, Aaron Aguirre, Junior Moura, Che Liu, Lanhai Zhong, Chenxi Sun, Gari Clifford, Brandon Westover, Shenda Hong

Artificial intelligence (AI) has demonstrated significant potential in ECG analysis and cardiovascular disease assessment. Recently, foundation models have played a remarkable role in advancing medical AI. The development of an ECG foundation model holds the promise of elevating AI-ECG research to new heights. However, building such a model faces several challenges, including insufficient database sample sizes and inadequate generalization across multiple domains. Additionally, there is a notable performance gap between single-lead and multi-lead ECG analyses. We introduced an ECG Foundation Model (ECGFounder), a general-purpose model that leverages real-world ECG annotations from cardiology experts to broaden the diagnostic capabilities of ECG analysis. ECGFounder was trained on over 10 million ECGs with 150 label categories from the Harvard-Emory ECG Database, enabling comprehensive cardiovascular disease diagnosis through ECG analysis. The model is designed to be both an effective out-of-the-box solution, and a to be fine-tunable for downstream tasks, maximizing usability. Importantly, we extended its application to lower rank ECGs, and arbitrary single-lead ECGs in particular. ECGFounder is applicable to supporting various downstream tasks in mobile monitoring scenarios. Experimental results demonstrate that ECGFounder achieves expert-level performance on internal validation sets, with AUROC exceeding 0.95 for eighty diagnoses. It also shows strong classification performance and generalization across various diagnoses on external validation sets. When fine-tuned, ECGFounder outperforms baseline models in demographic analysis, clinical event detection, and cross-modality cardiac rhythm diagnosis. The trained model and data will be publicly released upon publication through the bdsp.io. Our code is available at https://github.com/bdsp-core/ECGFounder

comment: working in progress

♻ ★ Any2Point: Empowering Any-modality Large Models for Efficient 3D Understanding

Yiwen Tang, Ray Zhang, Jiaming Liu, Zoey Guo, Dong Wang, Zhigang Wang, Bin Zhao, Shanghang Zhang, Peng Gao, Hongsheng Li, Xuelong Li

Large foundation models have recently emerged as a prominent focus of interest, attaining superior performance in widespread scenarios. Due to the scarcity of 3D data, many efforts have been made to adapt pre-trained transformers from vision to 3D domains. However, such 2D-to-3D approaches are still limited, due to the potential loss of spatial geometries and high computation cost. More importantly, their frameworks are mainly designed for 2D models, lacking a general any-to-3D paradigm. In this paper, we introduce Any2Point, a parameter-efficient method to empower any-modality large models (vision, language, audio) for 3D understanding. Given a frozen transformer from any source modality, we propose a 3D-to-any (1D or 2D) virtual projection strategy that correlates the input 3D points to the original 1D or 2D positions within the source modality. This mechanism enables us to assign each 3D token with a positional encoding paired with the pre-trained model, which avoids 3D geometry loss caused by the true projection and better motivates the transformer for 3D learning with 1D/2D positional priors. Then, within each transformer block, we insert an any-to-3D guided adapter module for parameter-efficient fine-tuning. The adapter incorporates prior spatial knowledge from the source modality to guide the local feature aggregation of 3D tokens, compelling the semantic adaption of any-modality transformers. We conduct extensive experiments to showcase the effectiveness and efficiency of our method. Code and models are released at https://github.com/Ivan-Tang-3D/Any2Point.

comment: Code and models are released at https://github.com/Ivan-Tang-3D/Any2Point

♻ ☆ Point-PEFT: Parameter-Efficient Fine-Tuning for 3D Pre-trained Models

Yiwen Tang, Ray Zhang, Zoey Guo, Dong Wang, Zhigang Wang, Bin Zhao, Xuelong Li

The popularity of pre-trained large models has revolutionized downstream tasks across diverse fields, such as language, vision, and multi-modality. To minimize the adaption cost for downstream tasks, many Parameter-Efficient Fine-Tuning (PEFT) techniques are proposed for language and 2D image pre-trained models. However, the specialized PEFT method for 3D pre-trained models is still under-explored. To this end, we introduce Point-PEFT, a novel framework for adapting point cloud pre-trained models with minimal learnable parameters. Specifically, for a pre-trained 3D model, we freeze most of its parameters, and only tune the newly added PEFT modules on downstream tasks, which consist of a Point-prior Prompt and a Geometry-aware Adapter. The Point-prior Prompt adopts a set of learnable prompt tokens, for which we propose to construct a memory bank with domain-specific knowledge, and utilize a parameter-free attention to enhance the prompt tokens. The Geometry-aware Adapter aims to aggregate point cloud features within spatial neighborhoods to capture fine-grained geometric information through local interactions. Extensive experiments indicate that our Point-PEFT can achieve better performance than the full fine-tuning on various downstream tasks, while using only 5% of the trainable parameters, demonstrating the efficiency and effectiveness of our approach. Code is released at https://github.com/Ivan-Tang-3D/Point-PEFT.

comment: The specialized PEFT framework for 3D pre-trained models, which achieves competitive performance to full fine-tuning, and significantly reduces the computational resources. Project page: https://github.com/Ivan-Tang-3D/Point-PEFT

♻ ☆ Amortized Planning with Large-Scale Transformers: A Case Study on Chess

Anian Ruoss, Grégoire Delétang, Sourabh Medapati, Jordi Grau-Moya, Li Kevin Wenliang, Elliot Catt, John Reid, Cannada A. Lewis, Joel Veness, Tim Genewein

This paper uses chess, a landmark planning problem in AI, to assess transformers' performance on a planning task where memorization is futile $\unicode{x2013}$ even at a large scale. To this end, we release ChessBench, a large-scale benchmark dataset of 10 million chess games with legal move and value annotations (15 billion data points) provided by Stockfish 16, the state-of-the-art chess engine. We train transformers with up to 270 million parameters on ChessBench via supervised learning and perform extensive ablations to assess the impact of dataset size, model size, architecture type, and different prediction targets (state-values, action-values, and behavioral cloning). Our largest models learn to predict action-values for novel boards quite accurately, implying highly non-trivial generalization. Despite performing no explicit search, our resulting chess policy solves challenging chess puzzles and achieves a surprisingly strong Lichess blitz Elo of 2895 against humans (grandmaster level). We also compare to Leela Chess Zero and AlphaZero (trained without supervision via self-play) with and without search. We show that, although a remarkably good approximation of Stockfish's search-based algorithm can be distilled into large-scale transformers via supervised learning, perfect distillation is still beyond reach, thus making ChessBench well-suited for future research.

♻ ☆ Deep Multimodal Learning with Missing Modality: A Survey

Renjie Wu, Hu Wang, Hsiang-Ting Chen, Gustavo Carneiro

During multimodal model training and testing, certain data modalities may be absent due to sensor limitations, cost constraints, privacy concerns, or data loss, negatively affecting performance. Multimodal learning techniques designed to handle missing modalities can mitigate this by ensuring model robustness even when some modalities are unavailable. This survey reviews recent progress in Multimodal Learning with Missing Modality (MLMM), focusing on deep learning methods. It provides the first comprehensive survey that covers the motivation and distinctions between MLMM and standard multimodal learning setups, followed by a detailed analysis of current methods, applications, and datasets, concluding with challenges and future directions.

comment: Submitted to ACM Computing Surveys

♻ ☆ LLM-based SPARQL Query Generation from Natural Language over Federated Knowledge Graphs

Vincent Emonet, Jerven Bolleman, Severine Duvaud, Tarcisio Mendes de Farias, Ana Claudia Sima

We introduce a Retrieval-Augmented Generation (RAG) system for translating user questions into accurate federated SPARQL queries over bioinformatics knowledge graphs (KGs) leveraging Large Language Models (LLMs). To enhance accuracy and reduce hallucinations in query generation, our system utilises metadata from the KGs, including query examples and schema information, and incorporates a validation step to correct generated queries. The system is available online at chat.expasy.org.

♻ ☆ Truth is Universal: Robust Detection of Lies in LLMs NeurIPS 2024

Lennart Bürger, Fred A. Hamprecht, Boaz Nadler

Large Language Models (LLMs) have revolutionised natural language processing, exhibiting impressive human-like capabilities. In particular, LLMs are capable of "lying", knowingly outputting false statements. Hence, it is of interest and importance to develop methods to detect when LLMs lie. Indeed, several authors trained classifiers to detect LLM lies based on their internal model activations. However, other researchers showed that these classifiers may fail to generalise, for example to negated statements. In this work, we aim to develop a robust method to detect when an LLM is lying. To this end, we make the following key contributions: (i) We demonstrate the existence of a two-dimensional subspace, along which the activation vectors of true and false statements can be separated. Notably, this finding is universal and holds for various LLMs, including Gemma-7B, LLaMA2-13B, Mistral-7B and LLaMA3-8B. Our analysis explains the generalisation failures observed in previous studies and sets the stage for more robust lie detection; (ii) Building upon (i), we construct an accurate LLM lie detector. Empirically, our proposed classifier achieves state-of-the-art performance, attaining 94% accuracy in both distinguishing true from false factual statements and detecting lies generated in real-world scenarios.

comment: NeurIPS 2024 poster

♻ ☆ On the token distance modeling ability of higher RoPE attention dimension EMNLP 2024

Xiangyu Hong, Che Jiang, Biqing Qi, Fandong Meng, Mo Yu, Bowen Zhou, Jie Zhou

Length extrapolation algorithms based on Rotary position embedding (RoPE) have shown promising results in extending the context length of language models. However, understanding how position embedding can capture longer-range contextual information remains elusive. Based on the intuition that different dimensions correspond to different frequency of changes in RoPE encoding, we conducted a dimension-level analysis to investigate the correlation between a hidden dimension of an attention head and its contribution to capturing long-distance dependencies. Using our correlation metric, we identified a particular type of attention heads, which we named Positional Heads, from various length-extrapolated models. These heads exhibit a strong focus on long-range information interaction and play a pivotal role in long input processing, as evidence by our ablation. We further demonstrate the correlation between the efficiency of length extrapolation and the extension of the high-dimensional attention allocation of these heads. The identification of Positional Heads provides insights for future research in long-text comprehension.

comment: Accepted to EMNLP 2024 Findings

♻ ☆ Why Transformers Need Adam: A Hessian Perspective

Yushun Zhang, Congliang Chen, Tian Ding, Ziniu Li, Ruoyu Sun, Zhi-Quan Luo

SGD performs worse than Adam by a significant margin on Transformers, but the reason remains unclear. In this work, we provide an explanation through the lens of Hessian: (i) Transformers are "heterogeneous": the Hessian spectrum across parameter blocks vary dramatically, a phenomenon we call "block heterogeneity"; (ii) Heterogeneity hampers SGD: SGD performs worse than Adam on problems with block heterogeneity. To validate (i) and (ii), we check various Transformers, CNNs, MLPs, and quadratic problems, and find that SGD can perform on par with Adam on problems without block heterogeneity, but performs worse than Adam when the heterogeneity exists. Our initial theoretical analysis indicates that SGD performs worse because it applies one single learning rate to all blocks, which cannot handle the heterogeneity among blocks. This limitation could be ameliorated if we use coordinate-wise learning rates, as designed in Adam.

comment: Advances in Neural Information Processing Systems, 2024

♻ ☆ QUIS: Question-guided Insights Generation for Automated Exploratory Data Analysis EMNLP 2024

Abhijit Manatkar, Ashlesha Akella, Parthivi Gupta, Krishnasuri Narayanam

Discovering meaningful insights from a large dataset, known as Exploratory Data Analysis (EDA), is a challenging task that requires thorough exploration and analysis of the data. Automated Data Exploration (ADE) systems use goal-oriented methods with Large Language Models and Reinforcement Learning towards full automation. However, these methods require human involvement to anticipate goals that may limit insight extraction, while fully automated systems demand significant computational resources and retraining for new datasets. We introduce QUIS, a fully automated EDA system that operates in two stages: insight generation (ISGen) driven by question generation (QUGen). The QUGen module generates questions in iterations, refining them from previous iterations to enhance coverage without human intervention or manually curated examples. The ISGen module analyzes data to produce multiple relevant insights in response to each question, requiring no prior training and enabling QUIS to adapt to new datasets.

comment: Accepted for EMNLP 2024 Industry Track

♻ ☆ Deconstructing The Ethics of Large Language Models from Long-standing Issues to New-emerging Dilemmas: A Survey

Chengyuan Deng, Yiqun Duan, Xin Jin, Heng Chang, Yijun Tian, Han Liu, Yichen Wang, Kuofeng Gao, Henry Peng Zou, Yiqiao Jin, Yijia Xiao, Shenghao Wu, Zongxing Xie, Weimin Lyu, Sihong He, Lu Cheng, Haohan Wang, Jun Zhuang

Large Language Models (LLMs) have achieved unparalleled success across diverse language modeling tasks in recent years. However, this progress has also intensified ethical concerns, impacting the deployment of LLMs in everyday contexts. This paper provides a comprehensive survey of ethical challenges associated with LLMs, from longstanding issues such as copyright infringement, systematic bias, and data privacy, to emerging problems like truthfulness and social norms. We critically analyze existing research aimed at understanding, examining, and mitigating these ethical risks. Our survey underscores integrating ethical standards and societal values into the development of LLMs, thereby guiding the development of responsible and ethically aligned language models.

♻ ☆ Can LLMs Recognize Toxicity? A Structured Investigation Framework and Toxicity Metric

Hyukhun Koh, Dohyung Kim, Minwoo Lee, Kyomin Jung

In the pursuit of developing Large Language Models (LLMs) that adhere to societal standards, it is imperative to detect the toxicity in the generated text. The majority of existing toxicity metrics rely on encoder models trained on specific toxicity datasets, which are susceptible to out-of-distribution (OOD) problems and depend on the dataset's definition of toxicity. In this paper, we introduce a robust metric grounded on LLMs to flexibly measure toxicity according to the given definition. We first analyze the toxicity factors, followed by an examination of the intrinsic toxic attributes of LLMs to ascertain their suitability as evaluators. Finally, we evaluate the performance of our metric with detailed analysis. Our empirical results demonstrate outstanding performance in measuring toxicity within verified factors, improving on conventional metrics by 12 points in the F1 score. Our findings also indicate that upstream toxicity significantly influences downstream metrics, suggesting that LLMs are unsuitable for toxicity evaluations within unverified factors.

comment: 8 page long

♻ ☆ NutrifyAI: An AI-Powered System for Real-Time Food Detection, Nutritional Analysis, and Personalized Meal Recommendations

Michelle Han, Junyao Chen, Zhengyuan Zhou

With diet and nutrition apps reaching 1.4 billion users in 2022 [1], it's not surprise that popular health apps, MyFitnessPal, Noom, and Calorie Counter, are surging in popularity. However, one major setback [2] of nearly all nutrition applications is that users must enter food data manually, which is time-consuming and tedious. Thus, there has been an increasing demand for applications that can accurately identify food items, analyze their nutritional content, and offer dietary recommendations in real-time. This paper introduces a comprehensive system that combines advanced computer vision techniques with nutritional analysis, implemented in a versatile mobile and web application. The system is divided into three key concepts: 1) food detection using the YOLOv8 model, 2) nutrient analysis via the Edamam Nutrition Analysis API, and 3) personalized meal recommendations using the Edamam Meal Planning and Recipe Search APIs. Preliminary results showcase the system's effectiveness by providing immediate, accurate dietary insights, with a demonstrated food recognition accuracy of nearly 80%, making it a valuable tool for users to make informed dietary decisions.

comment: 4 pages, 8 figures

♻ ☆ HiRT: Enhancing Robotic Control with Hierarchical Robot Transformers

Jianke Zhang, Yanjiang Guo, Xiaoyu Chen, Yen-Jen Wang, Yucheng Hu, Chengming Shi, Jianyu Chen

Large Vision-Language-Action (VLA) models, leveraging powerful pre trained Vision-Language Models (VLMs) backends, have shown promise in robotic control due to their impressive generalization ability. However, the success comes at a cost. Their reliance on VLM backends with billions of parameters leads to high computational costs and inference latency, limiting the testing scenarios to mainly quasi-static tasks and hindering performance in dynamic tasks requiring rapid interactions. To address these limitations, this paper proposes HiRT, a Hierarchical Robot Transformer framework that enables flexible frequency and performance trade-off. HiRT keeps VLMs running at low frequencies to capture temporarily invariant features while enabling real-time interaction through a high-frequency vision-based policy guided by the slowly updated features. Experiment results in both simulation and real-world settings demonstrate significant improvements over baseline methods. Empirically, in static tasks, we double the control frequency and achieve comparable success rates. Additionally, on novel real-world dynamic ma nipulation tasks which are challenging for previous VLA models, HiRT improves the success rate from 48% to 75%.

♻ ☆ Reinforcement Learning for Molecular Dynamics Optimization: A Stochastic Pontryagin Maximum Principle Approach ICONIP

Chandrajit Bajaj, Minh Nguyen, Conrad Li

In this paper, we present a novel reinforcement learning framework designed to optimize molecular dynamics by focusing on the entire trajectory rather than just the final molecular configuration. Leveraging a stochastic version of Pontryagin's Maximum Principle (PMP) and Soft Actor-Critic (SAC) algorithm, our framework effectively explores non-convex molecular energy landscapes, escaping local minima to stabilize in low-energy states. Our approach operates in continuous state and action spaces without relying on labeled data, making it applicable to a wide range of molecular systems. Through extensive experimentation on six distinct molecules, including Bradykinin and Oxytocin, we demonstrate competitive performance against other unsupervised physics-based methods, such as the Greedy and NEMO-based algorithms. Our method's adaptability and focus on dynamic trajectory optimization make it suitable for applications in areas such as drug discovery and molecular design.

comment: Accepted to the International Conference on Neural Information Processing (ICONIP) 2024. To be published in Springer-Nature Communications in Computer and Information Science (CCIS) Series

♻ ☆ Self-Supervised Learning of Disentangled Representations for Multivariate Time-Series NeurIPS

Ching Chang, Chiao-Tung Chan, Wei-Yao Wang, Wen-Chih Peng, Tien-Fu Chen

Multivariate time-series data in fields like healthcare and industry are informative but challenging due to high dimensionality and lack of labels. Recent self-supervised learning methods excel in learning rich representations without labels but struggle with disentangled embeddings and inductive bias issues like transformation-invariance. To address these challenges, we introduce TimeDRL, a framework for multivariate time-series representation learning with dual-level disentangled embeddings. TimeDRL features: (i) disentangled timestamp-level and instance-level embeddings using a [CLS] token strategy; (ii) timestamp-predictive and instance-contrastive tasks for representation learning; and (iii) avoidance of augmentation methods to eliminate inductive biases. Experiments on forecasting and classification datasets show TimeDRL outperforms existing methods, with further validation in semi-supervised settings with limited labeled data.

comment: This submission has been withdrawn to avoid duplication with a full version of the paper that is already available in another arXiv entry (arXiv:2410.12606). The withdrawn version was a short format prepared for a NeurIPS workshop and is no longer necessary as a separate arXiv submission

♻ ☆ Cardiac Copilot: Automatic Probe Guidance for Echocardiography with World Model MICCAI2024

Haojun Jiang, Zhenguo Sun, Ning Jia, Meng Li, Yu Sun, Shaqi Luo, Shiji Song, Gao Huang

Echocardiography is the only technique capable of real-time imaging of the heart and is vital for diagnosing the majority of cardiac diseases. However, there is a severe shortage of experienced cardiac sonographers, due to the heart's complex structure and significant operational challenges. To mitigate this situation, we present a Cardiac Copilot system capable of providing real-time probe movement guidance to assist less experienced sonographers in conducting freehand echocardiography. This system can enable non-experts, especially in primary departments and medically underserved areas, to perform cardiac ultrasound examinations, potentially improving global healthcare delivery. The core innovation lies in proposing a data-driven world model, named Cardiac Dreamer, for representing cardiac spatial structures. This world model can provide structure features of any cardiac planes around the current probe position in the latent space, serving as an precise navigation map for autonomous plane localization. We train our model with real-world ultrasound data and corresponding probe motion from 110 routine clinical scans with 151K sample pairs by three certified sonographers. Evaluations on three standard planes with 37K sample pairs demonstrate that the world model can reduce navigation errors by up to 33\% and exhibit more stable performance.

comment: Accepted by MICCAI2024

♻ ☆ FAMOUS: Flexible Accelerator for the Attention Mechanism of Transformer on UltraScale+ FPGAs

Ehsan Kabir, Md. Arafat Kabir, Austin R. J. Downey, Jason D. Bakos, David Andrews, Miaoqing Huang

Transformer neural networks (TNNs) are being applied across a widening range of application domains, including natural language processing (NLP), machine translation, and computer vision (CV). Their popularity is largely attributed to the exceptional performance of their multi-head self-attention blocks when analyzing sequential data and extracting features. To date, there are limited hardware accelerators tailored for this mechanism, which is the first step before designing an accelerator for a complete model. This paper proposes \textit{FAMOUS}, a flexible hardware accelerator for dense multi-head attention (MHA) computation of TNNs on field-programmable gate arrays (FPGAs). It is optimized for high utilization of processing elements and on-chip memories to improve parallelism and reduce latency. An efficient tiling of large matrices has been employed to distribute memory and computing resources across different modules on various FPGA platforms. The design is evaluated on Xilinx Alveo U55C and U200 data center cards containing Ultrascale+ FPGAs. Experimental results are presented that show that it can attain a maximum throughput, number of parallel attention heads, embedding dimension and tile size of 328 (giga operations/second (GOPS)), 8, 768 and 64 respectively on the U55C. Furthermore, it is 3.28$\times$ and 2.6$\times$ faster than the Intel Xeon Gold 5220R CPU and NVIDIA V100 GPU respectively. It is also 1.3$\times$ faster than the fastest state-of-the-art FPGA-based accelerator.

comment: arXiv admin note: text overlap with arXiv:2409.13975

♻ ☆ Trust or Bust: Ensuring Trustworthiness in Autonomous Weapon Systems

Kasper Cools, Clara Maathuis

The integration of Autonomous Weapon Systems (AWS) into military operations presents both significant opportunities and challenges. This paper explores the multifaceted nature of trust in AWS, emphasising the necessity of establishing reliable and transparent systems to mitigate risks associated with bias, operational failures, and accountability. Despite advancements in Artificial Intelligence (AI), the trustworthiness of these systems, especially in high-stakes military applications, remains a critical issue. Through a systematic review of existing literature, this research identifies gaps in the understanding of trust dynamics during the development and deployment phases of AWS. It advocates for a collaborative approach that includes technologists, ethicists, and military strategists to address these ongoing challenges. The findings underscore the importance of Human-Machine teaming and enhancing system intelligibility to ensure accountability and adherence to International Humanitarian Law. Ultimately, this paper aims to contribute to the ongoing discourse on the ethical implications of AWS and the imperative for trustworthy AI in defense contexts.

comment: Accepted as a workshop paper at MILCOM 2024, 8 pages

♻ ☆ FSL-Rectifier: Rectify Outliers in Few-Shot Learning via Test-Time Augmentation

Yunwei Bai, Ying Kiat Tan, Shiming Chen, Yao Shu, Tsuhan Chen

Few-shot-learning (FSL) commonly requires a model to identify images (queries) that belong to classes unseen during training, based on a few labeled samples of the new classes (support set) as reference. So far, plenty of algorithms involve training data augmentation to improve the generalization capability of FSL models, but outlier queries or support images during inference can still pose great generalization challenges. In this work, to reduce the bias caused by the outlier samples, we generate additional test-class samples by combining original samples with suitable train-class samples via a generative image combiner. Then, we obtain averaged features via an augmentor, which leads to more typical representations through the averaging. We experimentally and theoretically demonstrate the effectiveness of our method, e.g., obtaining a test accuracy improvement proportion of around 10% (e.g., from 46.86% to 53.28%) for trained FSL models. Importantly, given pretrained image combiner, our method is training-free for off-the-shelf FSL models, whose performance can be improved without extra datasets nor further training of the models themselves.

♻ ☆ Diffusion-TS: Interpretable Diffusion for General Time Series Generation

Xinyu Yuan, Yan Qiao

Denoising diffusion probabilistic models (DDPMs) are becoming the leading paradigm for generative models. It has recently shown breakthroughs in audio synthesis, time series imputation and forecasting. In this paper, we propose Diffusion-TS, a novel diffusion-based framework that generates multivariate time series samples of high quality by using an encoder-decoder transformer with disentangled temporal representations, in which the decomposition technique guides Diffusion-TS to capture the semantic meaning of time series while transformers mine detailed sequential information from the noisy model input. Different from existing diffusion-based approaches, we train the model to directly reconstruct the sample instead of the noise in each diffusion step, combining a Fourier-based loss term. Diffusion-TS is expected to generate time series satisfying both interpretablity and realness. In addition, it is shown that the proposed Diffusion-TS can be easily extended to conditional generation tasks, such as forecasting and imputation, without any model changes. This also motivates us to further explore the performance of Diffusion-TS under irregular settings. Finally, through qualitative and quantitative experiments, results show that Diffusion-TS achieves the state-of-the-art results on various realistic analyses of time series.

♻ ☆ AlphaEdit: Null-Space Constrained Knowledge Editing for Language Models

Junfeng Fang, Houcheng Jiang, Kun Wang, Yunshan Ma, Xiang Wang, Xiangnan He, Tat-seng Chua

Large language models (LLMs) often exhibit hallucinations due to incorrect or outdated knowledge. Hence, model editing methods have emerged to enable targeted knowledge updates. To achieve this, a prevailing paradigm is the locating-then-editing approach, which first locates influential parameters and then edits them by introducing a perturbation. While effective, current studies have demonstrated that this perturbation inevitably disrupt the originally preserved knowledge within LLMs, especially in sequential editing scenarios. To address this, we introduce AlphaEdit, a novel solution that projects perturbation onto the null space of the preserved knowledge before applying it to the parameters. We theoretically prove that this projection ensures the output of post-edited LLMs remains unchanged when queried about the preserved knowledge, thereby mitigating the issue of disruption. Extensive experiments on various LLMs, including LLaMA3, GPT2-XL, and GPT-J, show that AlphaEdit boosts the performance of most locating-then-editing methods by an average of 36.4% with a single line of additional code for projection solely. Our code is available at: https://github.com/jianghoucheng/AlphaEdit.

♻ ☆ Exploring the Zero-Shot Capabilities of LLMs Handling Multiple Problems at once

Zhengxiang Wang, Jordan Kodner, Owen Rambow

Recent studies have proposed placing multiple problems in a single prompt to improve input token utilization for a more efficient LLM inference. We call this MPP, in contrast to conventional SPP that prompts an LLM with a single problem at a time. While MPP has been shown to work comparably well or even better than SPP under few-shot settings, its zero-shot performance is underexplored, which better reveals the innate multiple problem handling capabilities of LLMs. To address that, we study the zero-shot MPP performance of various LLMs on 6 classification and 12 reasoning benchmarks and confirm that LLMs are competent zero-shot multi-problem solvers. We also examine the conditions of effectiveness of zero-shot MPP and explore several model-level factors that may enable MPP. We observe that LLMs consistently perform worse with selecting indices of texts of a given class label and with multiple mixed-source reasoning problems, indicating a lack of true understanding. We also find that instruction tuning is an important factor than enhances MPP.

comment: 26 pages, 11 figures, 16 tables

♻ ☆ Diffusion Models as Constrained Samplers for Optimization with Unknown Constraints

Lingkai Kong, Yuanqi Du, Wenhao Mu, Kirill Neklyudov, Valentin De Bortoli, Dongxia Wu, Haorui Wang, Aaron Ferber, Yi-An Ma, Carla P. Gomes, Chao Zhang

Addressing real-world optimization problems becomes particularly challenging when analytic objective functions or constraints are unavailable. While numerous studies have addressed the issue of unknown objectives, limited research has focused on scenarios where feasibility constraints are not given explicitly. Overlooking these constraints can lead to spurious solutions that are unrealistic in practice. To deal with such unknown constraints, we propose to perform optimization within the data manifold using diffusion models. To constrain the optimization process to the data manifold, we reformulate the original optimization problem as a sampling problem from the product of the Boltzmann distribution defined by the objective function and the data distribution learned by the diffusion model. Depending on the differentiability of the objective function, we propose two different sampling methods. For differentiable objectives, we propose a two-stage framework that begins with a guided diffusion process for warm-up, followed by a Langevin dynamics stage for further correction. For non-differentiable objectives, we propose an iterative importance sampling strategy using the diffusion model as the proposal distribution. Comprehensive experiments on a synthetic dataset, six real-world black-box optimization datasets, and a multi-objective molecule optimization dataset show that our method achieves better or comparable performance with previous state-of-the-art baselines.

♻ ☆ Open-World Continual Learning: Unifying Novelty Detection and Continual Learning

Gyuhak Kim, Changnan Xiao, Tatsuya Konishi, Zixuan Ke, Bing Liu

As AI agents are increasingly used in the real open world with unknowns or novelties, they need the ability to (1) recognize objects that (a) they have learned before and (b) detect items that they have never seen or learned, and (2) learn the new items incrementally to become more and more knowledgeable and powerful. (1) is called novelty detection or out-of-distribution (OOD) detection and (2) is called class incremental learning (CIL), which is a setting of continual learning (CL). In existing research, OOD detection and CIL are regarded as two completely different problems. This paper first provides a theoretical proof that good OOD detection for each task within the set of learned tasks (called closed-world OOD detection) is necessary for successful CIL. We show this by decomposing CIL into two sub-problems: within-task prediction (WP) and task-id prediction (TP), and proving that TP is correlated with closed-world OOD detection. The key theoretical result is that regardless of whether WP and OOD detection (or TP) are defined explicitly or implicitly by a CIL algorithm, good WP and good closed-world OOD detection are necessary and sufficient conditions for good CIL, which unifies novelty or OOD detection and continual learning (CIL, in particular). We call this traditional CIL the closed-world CIL as it does not detect future OOD data in the open world. The paper then proves that the theory can be generalized or extended to open-world CIL, which is the proposed open-world continual learning, that can perform CIL in the open world and detect future or open-world OOD data. Based on the theoretical results, new CIL methods are also designed, which outperform strong baselines in CIL accuracy and in continual OOD detection by a large margin.

comment: To appear in Artificial Intelligence Journal. arXiv admin note: substantial text overlap with arXiv:2211.02633

♻ ☆ The CLC-UKET Dataset: Benchmarking Case Outcome Prediction for the UK Employment Tribunal

Huiyuan Xie, Felix Steffek, Joana Ribeiro de Faria, Christine Carter, Jonathan Rutherford

This paper explores the intersection of technological innovation and access to justice by developing a benchmark for predicting case outcomes in the UK Employment Tribunal (UKET). To address the challenge of extensive manual annotation, the study employs a large language model (LLM) for automatic annotation, resulting in the creation of the CLC-UKET dataset. The dataset consists of approximately 19,000 UKET cases and their metadata. Comprehensive legal annotations cover facts, claims, precedent references, statutory references, case outcomes, reasons and jurisdiction codes. Facilitated by the CLC-UKET data, we examine a multi-class case outcome prediction task in the UKET. Human predictions are collected to establish a performance reference for model comparison. Empirical results from baseline models indicate that finetuned transformer models outperform zero-shot and few-shot LLMs on the UKET prediction task. The performance of zero-shot LLMs can be enhanced by integrating task-related information into few-shot examples. We hope that the CLC-UKET dataset, along with human annotations and empirical findings, can serve as a valuable benchmark for employment-related dispute resolution.

♻ ☆ PIR: Remote Sensing Image-Text Retrieval with Prior Instruction Representation Learning

Jiancheng Pan, Muyuan Ma, Qing Ma, Cong Bai, Shengyong Chen

Remote sensing image-text retrieval constitutes a foundational aspect of remote sensing interpretation tasks, facilitating the alignment of vision and language representations. This paper introduces a prior instruction representation (PIR) learning paradigm that draws on prior knowledge to instruct adaptive learning of vision and text representations. Based on PIR, a domain-adapted remote sensing image-text retrieval framework PIR-ITR is designed to address semantic noise issues in vision-language understanding tasks. However, with massive additional data for pre-training the vision-language foundation model, remote sensing image-text retrieval is further developed into an open-domain retrieval task. Continuing with the above, we propose PIR-CLIP, a domain-specific CLIP-based framework for remote sensing image-text retrieval, to address semantic noise in remote sensing vision-language representations and further improve open-domain retrieval performance. In vision representation, we utilize the prior-guided knowledge of the remote sensing scene recognition by building a belief matrix to select key features for reducing the impact of semantic noise. In text representation, we use the previous time step to cyclically activate the current time step to enhance text representation capability. A cluster-wise Affiliation Loss (AL) is proposed to constrain the inter-classes and to reduce the semantic confusion zones in the common subspace. Comprehensive experiments demonstrate that PIR could enhance vision and text representations and outperform the state-of-the-art methods of closed-domain and open-domain retrieval on two benchmark datasets, RSICD and RSITMD.

comment: 13 pages, 8 figures

♻ ☆ Chain of Ideas: Revolutionizing Research in Novel Idea Development with LLM Agents

Long Li, Weiwen Xu, Jiayan Guo, Ruochen Zhao, Xinxuan Li, Yuqian Yuan, Boqiang Zhang, Yuming Jiang, Yifei Xin, Ronghao Dang, Deli Zhao, Yu Rong, Tian Feng, Lidong Bing

Effective research ideation is a critical step for scientific research. However, the exponential increase in scientific literature makes it challenging for researchers to stay current with recent advances and identify meaningful research directions. Recent developments in large language models~(LLMs) suggest a promising avenue for automating the generation of novel research ideas. However, existing methods for idea generation either trivially prompt LLMs or directly expose LLMs to extensive literature without indicating useful information. Inspired by the research process of human researchers, we propose a Chain-of-Ideas~(CoI) agent, an LLM-based agent that organizes relevant literature in a chain structure to effectively mirror the progressive development in a research domain. This organization facilitates LLMs to capture the current advancements in research, thereby enhancing their ideation capabilities. Furthermore, we propose Idea Arena, an evaluation protocol that can comprehensively evaluate idea generation methods from different perspectives, aligning closely with the preferences of human researchers. Experimental results indicate that the CoI agent consistently outperforms other methods and shows comparable quality as humans in research idea generation. Moreover, our CoI agent is budget-friendly, with a minimum cost of \$0.50 to generate a candidate idea and its corresponding experimental design.

comment: 10 pages,5 figures, conference

♻ ☆ One2set + Large Language Model: Best Partners for Keyphrase Generation EMNLP 2024

Liangying Shao, Liang Zhang, Minlong Peng, Guoqi Ma, Hao Yue, Mingming Sun, Jinsong Su

Keyphrase generation (KPG) aims to automatically generate a collection of phrases representing the core concepts of a given document. The dominant paradigms in KPG include one2seq and one2set. Recently, there has been increasing interest in applying large language models (LLMs) to KPG. Our preliminary experiments reveal that it is challenging for a single model to excel in both recall and precision. Further analysis shows that: 1) the one2set paradigm owns the advantage of high recall, but suffers from improper assignments of supervision signals during training; 2) LLMs are powerful in keyphrase selection, but existing selection methods often make redundant selections. Given these observations, we introduce a generate-then-select framework decomposing KPG into two steps, where we adopt a one2set-based model as generator to produce candidates and then use an LLM as selector to select keyphrases from these candidates. Particularly, we make two important improvements on our generator and selector: 1) we design an Optimal Transport-based assignment strategy to address the above improper assignments; 2) we model the keyphrase selection as a sequence labeling task to alleviate redundant selections. Experimental results on multiple benchmark datasets show that our framework significantly surpasses state-of-the-art models, especially in absent keyphrase prediction.

comment: Accepted by EMNLP 2024 Main Conference

♻ ☆ Negative Sampling in Knowledge Graph Representation Learning: A Review

Tiroshan Madushanka, Ryutaro Ichise

Knowledge Graph Representation Learning (KGRL), or Knowledge Graph Embedding (KGE), is essential for AI applications such as knowledge construction and information retrieval. These models encode entities and relations into lower-dimensional vectors, supporting tasks like link prediction and recommendation systems. Training KGE models relies on both positive and negative samples for effective learning, but generating high-quality negative samples from existing knowledge graphs is challenging. The quality of these samples significantly impacts the model's accuracy. This comprehensive survey paper systematically reviews various negative sampling (NS) methods and their contributions to the success of KGRL. Their respective advantages and disadvantages are outlined by categorizing existing NS methods into six distinct categories. Moreover, this survey identifies open research questions that serve as potential directions for future investigations. By offering a generalization and alignment of fundamental NS concepts, this survey provides valuable insights for designing effective NS methods in the context of KGRL and serves as a motivating force for further advancements in the field.

♻ ☆ LW-FedSSL: Resource-efficient Layer-wise Federated Self-supervised Learning

Ye Lin Tun, Chu Myaet Thwal, Le Quang Huy, Minh N. H. Nguyen, Choong Seon Hong

Many studies integrate federated learning (FL) with self-supervised learning (SSL) to take advantage of raw data distributed across edge devices. However, edge devices often struggle with high computation and communication costs imposed by SSL and FL algorithms. To tackle this hindrance, we propose LW-FedSSL, a layer-wise federated self-supervised learning approach that allows edge devices to incrementally train a single layer of the model at a time. We introduce server-side calibration and representation alignment mechanisms to ensure LW-FedSSL delivers performance on par with conventional federated self-supervised learning (FedSSL) while significantly lowering resource demands. In a pure layer-wise training scheme, training one layer at a time may limit effective interaction between different layers of the model. The server-side calibration mechanism takes advantage of the resource-rich FL server to ensure smooth collaboration between different layers of the global model. During local training, the representation alignment mechanism encourages closeness between representations of local models and those of the global model, thereby preserving the layer cohesion established by server-side calibration. With the proposed mechanisms, LW-FedSSL achieves a $3.3 \times$ reduction in memory usage, $2.1 \times$ fewer computational operations (FLOPs), and a $3.2 \times$ lower communication cost while maintaining the same level of performance as its end-to-end training counterpart. Additionally, we explore a progressive training strategy called Prog-FedSSL, which matches end-to-end training in memory requirements but offers a $1.8 \times$ reduction in FLOPs and communication costs. Although Prog-FedSSL is not as resource-efficient as LW-FedSSL, its performance improvements make it a suitable candidate for FL environments with more lenient resource constraints.

♻ ☆ Stepwise Alignment for Constrained Language Model Policy Optimization NeurIPS 2024

Akifumi Wachi, Thien Q. Tran, Rei Sato, Takumi Tanabe, Youhei Akimoto

Safety and trustworthiness are indispensable requirements for real-world applications of AI systems using large language models (LLMs). This paper formulates human value alignment as an optimization problem of the language model policy to maximize reward under a safety constraint, and then proposes an algorithm, Stepwise Alignment for Constrained Policy Optimization (SACPO). One key idea behind SACPO, supported by theory, is that the optimal policy incorporating reward and safety can be directly obtained from a reward-aligned policy. Building on this key idea, SACPO aligns LLMs step-wise with each metric while leveraging simple yet powerful alignment algorithms such as direct preference optimization (DPO). SACPO offers several advantages, including simplicity, stability, computational efficiency, and flexibility of algorithms and datasets. Under mild assumptions, our theoretical analysis provides the upper bounds on optimality and safety constraint violation. Our experimental results show that SACPO can fine-tune Alpaca-7B better than the state-of-the-art method in terms of both helpfulness and harmlessness.

comment: Accepted at NeurIPS 2024. Code and models are available at https://github.com/line/sacpo

♻ ☆ Utilizing Large Language Models in An Iterative Paradigm with Domain Feedback for Molecule Optimization

Khiem Le, Nitesh V. Chawla

Molecule optimization is a critical task in drug discovery to optimize desired properties of a given molecule through chemical modification. Despite Large Language Models (LLMs) holding the potential to efficiently simulate this task by using natural language to direct the optimization, straightforwardly utilizing shows limited performance. In this work, we facilitate utilizing LLMs in an iterative paradigm by proposing a simple yet highly effective domain feedback provider, namely $\text{Re}^2$DF. In detail, $\text{Re}^2$DF harnesses an external toolkit, RDKit, to handle the molecule hallucination, if the modified molecule is chemically invalid. Otherwise, its desired properties are computed and compared to the original one, establishing reliable domain feedback with correct direction and distance towards the objective, followed by a retrieved example, to explicitly guide the LLM to refine the modified molecule. We conduct experiments across both single- and multi-property objectives with 2 thresholds, where $\text{Re}^2$DF shows significant improvements. Particularly, for 20 single-property objectives, $\text{Re}^2$DF enhances Hit ratio by 16.95% and 20.76% under loose and strict thresholds, respectively. For 32 multi-property objectives, $\text{Re}^2$DF enhances Hit ratio by 6.04% and 5.25%.

☆ xGen-MM-Vid (BLIP-3-Video): You Only Need 32 Tokens to Represent a Video Even in VLMs

Michael S. Ryoo, Honglu Zhou, Shrikant Kendre, Can Qin, Le Xue, Manli Shu, Silvio Savarese, Ran Xu, Caiming Xiong, Juan Carlos Niebles

We present xGen-MM-Vid (BLIP-3-Video): a multimodal language model for videos, particularly designed to efficiently capture temporal information over multiple frames. BLIP-3-Video takes advantage of the 'temporal encoder' in addition to the conventional visual tokenizer, which maps a sequence of tokens over multiple frames into a compact set of visual tokens. This enables BLIP3-Video to use much fewer visual tokens than its competing models (e.g., 32 vs. 4608 tokens). We explore different types of temporal encoders, including learnable spatio-temporal pooling as well as sequential models like Token Turing Machines. We experimentally confirm that BLIP-3-Video obtains video question-answering accuracies comparable to much larger state-of-the-art models (e.g., 34B), while being much smaller (i.e., 4B) and more efficient by using fewer visual tokens. The project website is at https://www.salesforceairesearch.com/opensource/xGen-MM-Vid/index.html

☆ Revisiting Deep Feature Reconstruction for Logical and Structural Industrial Anomaly Detection

Sukanya Patra, Souhaib Ben Taieb

Industrial anomaly detection is crucial for quality control and predictive maintenance, but it presents challenges due to limited training data, diverse anomaly types, and external factors that alter object appearances. Existing methods commonly detect structural anomalies, such as dents and scratches, by leveraging multi-scale features from image patches extracted through deep pre-trained networks. However, significant memory and computational demands often limit their practical application. Additionally, detecting logical anomalies-such as images with missing or excess elements-requires an understanding of spatial relationships that traditional patch-based methods fail to capture. In this work, we address these limitations by focusing on Deep Feature Reconstruction (DFR), a memory- and compute-efficient approach for detecting structural anomalies. We further enhance DFR into a unified framework, called ULSAD, which is capable of detecting both structural and logical anomalies. Specifically, we refine the DFR training objective to improve performance in structural anomaly detection, while introducing an attention-based loss mechanism using a global autoencoder-like network to handle logical anomaly detection. Our empirical evaluation across five benchmark datasets demonstrates the performance of ULSAD in detecting and localizing both structural and logical anomalies, outperforming eight state-of-the-art methods. An extensive ablation study further highlights the contribution of each component to the overall performance improvement. Our code is available at https://github.com/sukanyapatra1997/ULSAD-2024.git

comment: Accepted in Transactions on Machine Learning Research (TMLR). Link to OpenReview: https://openreview.net/forum?id=kdTC4ktHPD

☆ Distribution Learning with Valid Outputs Beyond the Worst-Case

Nick Rittler, Kamalika Chaudhuri

Generative models at times produce "invalid" outputs, such as images with generation artifacts and unnatural sounds. Validity-constrained distribution learning attempts to address this problem by requiring that the learned distribution have a provably small fraction of its mass in invalid parts of space -- something which standard loss minimization does not always ensure. To this end, a learner in this model can guide the learning via "validity queries", which allow it to ascertain the validity of individual examples. Prior work on this problem takes a worst-case stance, showing that proper learning requires an exponential number of validity queries, and demonstrating an improper algorithm which -- while generating guarantees in a wide-range of settings -- makes an atypical polynomial number of validity queries. In this work, we take a first step towards characterizing regimes where guaranteeing validity is easier than in the worst-case. We show that when the data distribution lies in the model class and the log-loss is minimized, the number of samples required to ensure validity has a weak dependence on the validity requirement. Additionally, we show that when the validity region belongs to a VC-class, a limited number of validity queries are often sufficient.

☆ Implicit Regularization for Tubal Tensor Factorizations via Gradient Descent

Santhosh Karnik, Anna Veselovska, Mark Iwen, Felix Krahmer

We provide a rigorous analysis of implicit regularization in an overparametrized tensor factorization problem beyond the lazy training regime. For matrix factorization problems, this phenomenon has been studied in a number of works. A particular challenge has been to design universal initialization strategies which provably lead to implicit regularization in gradient-descent methods. At the same time, it has been argued by Cohen et. al. 2016 that more general classes of neural networks can be captured by considering tensor factorizations. However, in the tensor case, implicit regularization has only been rigorously established for gradient flow or in the lazy training regime. In this paper, we prove the first tensor result of its kind for gradient descent rather than gradient flow. We focus on the tubal tensor product and the associated notion of low tubal rank, encouraged by the relevance of this model for image data. We establish that gradient descent in an overparametrized tensor factorization model with a small random initialization exhibits an implicit bias towards solutions of low tubal rank. Our theoretical findings are illustrated in an extensive set of numerical simulations show-casing the dynamics predicted by our theory as well as the crucial role of using a small random initialization.

comment: 58 pages, 4 figures

☆ MoRE: Multi-Modal Contrastive Pre-training with Transformers on X-Rays, ECGs, and Diagnostic Report

Samrajya Thapa, Koushik Howlader, Subhankar Bhattacharjee, Wei le

In this paper, we introduce a novel Multi-Modal Contrastive Pre-training Framework that synergistically combines X-rays, electrocardiograms (ECGs), and radiology/cardiology reports. Our approach leverages transformers to encode these diverse modalities into a unified representation space, aiming to enhance diagnostic accuracy and facilitate comprehensive patient assessments. We utilize LoRA-Peft to significantly reduce trainable parameters in the LLM and incorporate recent linear attention dropping strategy in the Vision Transformer(ViT) for smoother attention. Furthermore, we provide novel multimodal attention explanations and retrieval for our model. To the best of our knowledge, we are the first to propose an integrated model that combines X-ray, ECG, and Radiology/Cardiology Report with this approach. By utilizing contrastive loss, MoRE effectively aligns modality-specific features into a coherent embedding, which supports various downstream tasks such as zero-shot classification and multimodal retrieval. Employing our proposed methodology, we achieve state-of-the-art (SOTA) on the Mimic-IV, CheXpert, Edema Severity, and PtbXl downstream datasets, surpassing existing multimodal approaches. Our proposed framework shows significant improvements in capturing intricate inter-modal relationships and its robustness in medical diagnosis that establishes a framework for future research in multimodal learning in the healthcare sector.

comment: 10 pages, 5 figures, 9 tables. Supplementary detail in Appendix. Code made available in Github for reproducibility

☆ A Realistic Threat Model for Large Language Model Jailbreaks

Valentyn Boreiko, Alexander Panfilov, Vaclav Voracek, Matthias Hein, Jonas Geiping

A plethora of jailbreaking attacks have been proposed to obtain harmful responses from safety-tuned LLMs. In their original settings, these methods all largely succeed in coercing the target output, but their attacks vary substantially in fluency and computational effort. In this work, we propose a unified threat model for the principled comparison of these methods. Our threat model combines constraints in perplexity, measuring how far a jailbreak deviates from natural text, and computational budget, in total FLOPs. For the former, we build an N-gram model on 1T tokens, which, in contrast to model-based perplexity, allows for an LLM-agnostic and inherently interpretable evaluation. We adapt popular attacks to this new, realistic threat model, with which we, for the first time, benchmark these attacks on equal footing. After a rigorous comparison, we not only find attack success rates against safety-tuned modern models to be lower than previously presented but also find that attacks based on discrete optimization significantly outperform recent LLM-based attacks. Being inherently interpretable, our threat model allows for a comprehensive analysis and comparison of jailbreak attacks. We find that effective attacks exploit and abuse infrequent N-grams, either selecting N-grams absent from real-world text or rare ones, e.g. specific to code datasets.

☆ Comprehensive benchmarking of large language models for RNA secondary structure prediction

L. I. Zablocki, L. A. Bugnon, M. Gerard, L. Di Persia, G. Stegmayer, D. H. Milone

Inspired by the success of large language models (LLM) for DNA and proteins, several LLM for RNA have been developed recently. RNA-LLM uses large datasets of RNA sequences to learn, in a self-supervised way, how to represent each RNA base with a semantically rich numerical vector. This is done under the hypothesis that obtaining high-quality RNA representations can enhance data-costly downstream tasks. Among them, predicting the secondary structure is a fundamental task for uncovering RNA functional mechanisms. In this work we present a comprehensive experimental analysis of several pre-trained RNA-LLM, comparing them for the RNA secondary structure prediction task in an unified deep learning framework. The RNA-LLM were assessed with increasing generalization difficulty on benchmark datasets. Results showed that two LLM clearly outperform the other models, and revealed significant challenges for generalization in low-homology scenarios.

☆ Compute-Constrained Data Selection

Junjie Oscar Yin, Alexander M. Rush

Data selection can reduce the amount of training data needed to finetune LLMs; however, the efficacy of data selection scales directly with its compute. Motivated by the practical challenge of compute-constrained finetuning, we consider the setting in which both the cost of selecting data and training are budgeted for. We first formalize the problem of data selection with a cost-aware utility function, and model the data selection problem as trading off initial-selection cost for training gain. We run a comprehensive sweep of experiments across multiple tasks, varying compute budget by scaling finetuning tokens, model sizes, and data selection compute. These experiments show the validity of this model in real-world experiments. Interestingly we find that many powerful data selection methods are almost never compute-optimal, and that cheaper data selection alternatives dominate both from a theoretical and empirical perspective.

☆ CoT-TL: Low-Resource Temporal Knowledge Representation of Planning Instructions Using Chain-of-Thought Reasoning IROS 2024

Kumar Manas, Stefan Zwicklbauer, Adrian Paschke

Autonomous agents often face the challenge of interpreting uncertain natural language instructions for planning tasks. Representing these instructions as Linear Temporal Logic (LTL) enables planners to synthesize actionable plans. We introduce CoT-TL, a data-efficient in-context learning framework for translating natural language specifications into LTL representations. CoT-TL addresses the limitations of large language models, which typically rely on extensive fine-tuning data, by extending chain-of-thought reasoning and semantic roles to align with the requirements of formal logic creation. This approach enhances the transparency and rationale behind LTL generation, fostering user trust. CoT-TL achieves state-of-the-art accuracy across three diverse datasets in low-data scenarios, outperforming existing methods without fine-tuning or intermediate translations. To improve reliability and minimize hallucinations, we incorporate model checking to validate the syntax of the generated LTL output. We further demonstrate CoT-TL's effectiveness through ablation studies and evaluations on unseen LTL structures and formulas in a new dataset. Finally, we validate CoT-TL's practicality by integrating it into a QuadCopter for multi-step drone planning based on natural language instructions.

comment: Accepted for publication in Proceedings of the 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2024), Abu Dhabi 14-18 October 2024

☆ Systematic Review: Text Processing Algorithms in Machine Learning and Deep Learning for Mental Health Detection on Social Media

Yuchen Cao, Jianglai Dai, Zhongyan Wang, Yeyubei Zhang, Xiaorui Shen, Yunchong Liu, Yexin Tian

The global rise in depression necessitates innovative detection methods for early intervention. Social media provides a unique opportunity to identify depression through user-generated posts. This systematic review evaluates machine learning (ML) models for depression detection on social media, focusing on biases and methodological challenges throughout the ML lifecycle. A search of PubMed, IEEE Xplore, and Google Scholar identified 47 relevant studies published after 2010. The Prediction model Risk Of Bias ASsessment Tool (PROBAST) was utilized to assess methodological quality and risk of bias. Significant biases impacting model reliability and generalizability were found. There is a predominant reliance on Twitter (63.8%) and English-language content (over 90%), with most studies focusing on users from the United States and Europe. Non-probability sampling methods (approximately 80%) limit representativeness. Only 23% of studies explicitly addressed linguistic nuances like negations, crucial for accurate sentiment analysis. Inconsistent hyperparameter tuning was observed, with only 27.7% properly tuning models. About 17% did not adequately partition data into training, validation, and test sets, risking overfitting. While 74.5% used appropriate evaluation metrics for imbalanced data, others relied on accuracy without addressing class imbalance, potentially skewing results. Reporting transparency varied, often lacking critical methodological details. These findings highlight the need to diversify data sources, standardize preprocessing protocols, ensure consistent model development practices, address class imbalance, and enhance reporting transparency. By overcoming these challenges, future research can develop more robust and generalizable ML models for depression detection on social media, contributing to improved mental health outcomes globally.

☆ Theoretical Limitations of Ensembles in the Age of Overparameterization

Niclas Dern, John P. Cunningham, Geoff Pleiss

Classic tree-based ensembles generalize better than any single decision tree. In contrast, recent empirical studies find that modern ensembles of (overparameterized) neural networks may not provide any inherent generalization advantage over single but larger neural networks. This paper clarifies how modern overparameterized ensembles differ from their classic underparameterized counterparts, using ensembles of random feature (RF) regressors as a basis for developing theory. In contrast to the underparameterized regime, where ensembling typically induces regularization and increases generalization, we prove that infinite ensembles of overparameterized RF regressors become pointwise equivalent to (single) infinite-width RF regressors. This equivalence, which is exact for ridgeless models and approximate for small ridge penalties, implies that overparameterized ensembles and single large models exhibit nearly identical generalization. As a consequence, we can characterize the predictive variance amongst ensemble members, and demonstrate that it quantifies the expected effects of increasing capacity rather than capturing any conventional notion of uncertainty. Our results challenge common assumptions about the advantages of ensembles in overparameterized settings, prompting a reconsideration of how well intuitions from underparameterized ensembles transfer to deep ensembles and the overparameterized regime.

comment: 26 pages, 12 figures

☆ A Trust-Region Method for Graphical Stein Variational Inference

Liam Pavlovic, David M. Rosen

Stein variational inference (SVI) is a sample-based approximate Bayesian inference technique that generates a sample set by jointly optimizing the samples' locations to minimize an information-theoretic measure of discrepancy with the target probability distribution. SVI thus provides a fast and significantly more sample-efficient approach to Bayesian inference than traditional (random-sampling-based) alternatives. However, the optimization techniques employed in existing SVI methods struggle to address problems in which the target distribution is high-dimensional, poorly-conditioned, or non-convex, which severely limits the range of their practical applicability. In this paper, we propose a novel trust-region optimization approach for SVI that successfully addresses each of these challenges. Our method builds upon prior work in SVI by leveraging conditional independences in the target distribution (to achieve high-dimensional scaling) and second-order information (to address poor conditioning), while additionally providing an effective adaptive step control procedure, which is essential for ensuring convergence on challenging non-convex optimization problems. Experimental results show our method achieves superior numerical performance, both in convergence rate and sample accuracy, and scales better in high-dimensional distributions, than previous SVI techniques.

☆ MagicPIG: LSH Sampling for Efficient LLM Generation

Zhuoming Chen, Ranajoy Sadhukhan, Zihao Ye, Yang Zhou, Jianyu Zhang, Niklas Nolte, Yuandong Tian, Matthijs Douze, Leon Bottou, Zhihao Jia, Beidi Chen

Large language models (LLMs) with long context windows have gained significant attention. However, the KV cache, stored to avoid re-computation, becomes a bottleneck. Various dynamic sparse or TopK-based attention approximation methods have been proposed to leverage the common insight that attention is sparse. In this paper, we first show that TopK attention itself suffers from quality degradation in certain downstream tasks because attention is not always as sparse as expected. Rather than selecting the keys and values with the highest attention scores, sampling with theoretical guarantees can provide a better estimation for attention output. To make the sampling-based approximation practical in LLM generation, we propose MagicPIG, a heterogeneous system based on Locality Sensitive Hashing (LSH). MagicPIG significantly reduces the workload of attention computation while preserving high accuracy for diverse tasks. MagicPIG stores the LSH hash tables and runs the attention computation on the CPU, which allows it to serve longer contexts and larger batch sizes with high approximation accuracy. MagicPIG can improve decoding throughput by $1.9\sim3.9\times$ across various GPU hardware and achieve 110ms decoding latency on a single RTX 4090 for Llama-3.1-8B-Instruct model with a context of 96k tokens. The code is available at \url{https://github.com/Infini-AI-Lab/MagicPIG}.

☆ DMM: Distributed Matrix Mechanism for Differentially-Private Federated Learning using Packed Secret Sharing

Alexander Bienstock, Ujjwal Kumar, Antigoni Polychroniadou

Federated Learning (FL) has gained lots of traction recently, both in industry and academia. In FL, a machine learning model is trained using data from various end-users arranged in committees across several rounds. Since such data can often be sensitive, a primary challenge in FL is providing privacy while still retaining utility of the model. Differential Privacy (DP) has become the main measure of privacy in the FL setting. DP comes in two flavors: central and local. In the former, a centralized server is trusted to receive the users' raw gradients from a training step, and then perturb their aggregation with some noise before releasing the next version of the model. In the latter (more private) setting, noise is applied on users' local devices, and only the aggregation of users' noisy gradients is revealed even to the server. Great strides have been made in increasing the privacy-utility trade-off in the central DP setting, by utilizing the so-called matrix mechanism. However, progress has been mostly stalled in the local DP setting. In this work, we introduce the distributed matrix mechanism to achieve the best-of-both-worlds; local DP and also better privacy-utility trade-off from the matrix mechanism. We accomplish this by proposing a cryptographic protocol that securely transfers sensitive values across rounds, which makes use of packed secret sharing. This protocol accommodates the dynamic participation of users per training round required by FL, including those that may drop out from the computation. We provide experiments which show that our mechanism indeed significantly improves the privacy-utility trade-off of FL models compared to previous local DP mechanisms, with little added overhead.

☆ Metric as Transform: Exploring beyond Affine Transform for Interpretable Neural Network

Suman Sapkota

Artificial Neural Networks of varying architectures are generally paired with affine transformation at the core. However, we find dot product neurons with global influence less interpretable as compared to local influence of euclidean distance (as used in Radial Basis Function Network). In this work, we explore the generalization of dot product neurons to $l^p$-norm, metrics, and beyond. We find that metrics as transform performs similarly to affine transform when used in MultiLayer Perceptron or Convolutional Neural Network. Moreover, we explore various properties of Metrics, compare it with Affine, and present multiple cases where metrics seem to provide better interpretability. We develop an interpretable local dictionary based Neural Networks and use it to understand and reject adversarial examples.

comment: 22 pages, 20 figures, 3 tables

☆ Unsupervised Replay Strategies for Continual Learning with Limited Data

Anthony Bazhenov, Pahan Dewasurendra, Giri P. Krishnan, Jean Erik Delanois

Artificial neural networks (ANNs) show limited performance with scarce or imbalanced training data and face challenges with continuous learning, such as forgetting previously learned data after new tasks training. In contrast, the human brain can learn continuously and from just a few examples. This research explores the impact of 'sleep', an unsupervised phase incorporating stochastic activation with local Hebbian learning rules, on ANNs trained incrementally with limited and imbalanced datasets, specifically MNIST and Fashion MNIST. We discovered that introducing a sleep phase significantly enhanced accuracy in models trained with limited data. When a few tasks were trained sequentially, sleep replay not only rescued previously learned information that had been catastrophically forgetting following new task training but often enhanced performance in prior tasks, especially those trained with limited data. This study highlights the multifaceted role of sleep replay in augmenting learning efficiency and facilitating continual learning in ANNs.

☆ Warped Diffusion: Solving Video Inverse Problems with Image Diffusion Models NeurIPS 2024

Giannis Daras, Weili Nie, Karsten Kreis, Alex Dimakis, Morteza Mardani, Nikola Borislavov Kovachki, Arash Vahdat

Using image models naively for solving inverse video problems often suffers from flickering, texture-sticking, and temporal inconsistency in generated videos. To tackle these problems, in this paper, we view frames as continuous functions in the 2D space, and videos as a sequence of continuous warping transformations between different frames. This perspective allows us to train function space diffusion models only on images and utilize them to solve temporally correlated inverse problems. The function space diffusion models need to be equivariant with respect to the underlying spatial transformations. To ensure temporal consistency, we introduce a simple post-hoc test-time guidance towards (self)-equivariant solutions. Our method allows us to deploy state-of-the-art latent diffusion models such as Stable Diffusion XL to solve video inverse problems. We demonstrate the effectiveness of our method for video inpainting and $8\times$ video super-resolution, outperforming existing techniques based on noise transformations. We provide generated video results: https://giannisdaras.github.io/warped\_diffusion.github.io/.

comment: Accepted in NeurIPS 2024

☆ Small Contributions, Small Networks: Efficient Neural Network Pruning Based on Relative Importance

Mostafa Hussien, Mahmoud Afifi, Kim Khoa Nguyen, Mohamed Cheriet

Recent advancements have scaled neural networks to unprecedented sizes, achieving remarkable performance across a wide range of tasks. However, deploying these large-scale models on resource-constrained devices poses significant challenges due to substantial storage and computational requirements. Neural network pruning has emerged as an effective technique to mitigate these limitations by reducing model size and complexity. In this paper, we introduce an intuitive and interpretable pruning method based on activation statistics, rooted in information theory and statistical analysis. Our approach leverages the statistical properties of neuron activations to identify and remove weights with minimal contributions to neuron outputs. Specifically, we build a distribution of weight contributions across the dataset and utilize its parameters to guide the pruning process. Furthermore, we propose a Pruning-aware Training strategy that incorporates an additional regularization term to enhance the effectiveness of our pruning method. Extensive experiments on multiple datasets and network architectures demonstrate that our method consistently outperforms several baseline and state-of-the-art pruning techniques.

☆ Modelling Structured Data Learning with Restricted Boltzmann Machines in the Teacher-Student Setting

Robin Thériault, Francesco Tosello, Daniele Tantari

Restricted Boltzmann machines (RBM) are generative models capable to learn data with a rich underlying structure. We study the teacher-student setting where a student RBM learns structured data generated by a teacher RBM. The amount of structure in the data is controlled by adjusting the number of hidden units of the teacher and the correlations in the rows of the weights, a.k.a. patterns. In the absence of correlations, we validate the conjecture that the performance is independent of the number of teacher patters and hidden units of the student RBMs, and we argue that the teacher-student setting can be used as a toy model for studying the lottery ticket hypothesis. Beyond this regime, we find that the critical amount of data required to learn the teacher patterns decreases with both their number and correlations. In both regimes, we find that, even with an relatively large dataset, it becomes impossible to learn the teacher patterns if the inference temperature used for regularization is kept too low. In our framework, the student can learn teacher patterns one-to-one or many-to-one, generalizing previous findings about the teacher-student setting with two hidden units to any arbitrary finite number of hidden units.

comment: 51 pages, 21 figures

☆ Towards Combating Frequency Simplicity-biased Learning for Domain Generalization NeurIPS 2024

Xilin He, Jingyu Hu, Qinliang Lin, Cheng Luo, Weicheng Xie, Siyang Song, Muhammad Haris Khan, Linlin Shen

Domain generalization methods aim to learn transferable knowledge from source domains that can generalize well to unseen target domains. Recent studies show that neural networks frequently suffer from a simplicity-biased learning behavior which leads to over-reliance on specific frequency sets, namely as frequency shortcuts, instead of semantic information, resulting in poor generalization performance. Despite previous data augmentation techniques successfully enhancing generalization performances, they intend to apply more frequency shortcuts, thereby causing hallucinations of generalization improvement. In this paper, we aim to prevent such learning behavior of applying frequency shortcuts from a data-driven perspective. Given the theoretical justification of models' biased learning behavior on different spatial frequency components, which is based on the dataset frequency properties, we argue that the learning behavior on various frequency components could be manipulated by changing the dataset statistical structure in the Fourier domain. Intuitively, as frequency shortcuts are hidden in the dominant and highly dependent frequencies of dataset structure, dynamically perturbating the over-reliance frequency components could prevent the application of frequency shortcuts. To this end, we propose two effective data augmentation modules designed to collaboratively and adaptively adjust the frequency characteristic of the dataset, aiming to dynamically influence the learning behavior of the model and ultimately serving as a strategy to mitigate shortcut learning. Code is available at AdvFrequency (https://github.com/C0notSilly/AdvFrequency).

comment: Accepted by NeurIPS 2024

☆ Theoretical Insights into Line Graph Transformation on Graph Learning

Fan Yang, Xingyue Huang

Line graph transformation has been widely studied in graph theory, where each node in a line graph corresponds to an edge in the original graph. This has inspired a series of graph neural networks (GNNs) applied to transformed line graphs, which have proven effective in various graph representation learning tasks. However, there is limited theoretical study on how line graph transformation affects the expressivity of GNN models. In this study, we focus on two types of graphs known to be challenging to the Weisfeiler-Leman (WL) tests: Cai-F\"urer-Immerman (CFI) graphs and strongly regular graphs, and show that applying line graph transformation helps exclude these challenging graph properties, thus potentially assist WL tests in distinguishing these graphs. We empirically validate our findings by conducting a series of experiments that compare the accuracy and efficiency of graph isomorphism tests and GNNs on both line-transformed and original graphs across these graph structure types.

comment: 21 pages, code available at https://github.com/lukeyf/graphs-and-lines

☆ Beyond 2:4: exploring V:N:M sparsity for efficient transformer inference on GPUs

Kang Zhao, Tao Yuan, Han Bao, Zhenfeng Su, Chang Gao, Zhaofeng Sun, Zichen Liang, Liping Jing, Jianfei Chen

To date, 2:4 sparsity has stood as the only sparse pattern that can be accelerated using sparse tensor cores on GPUs. In practice, 2:4 sparsity often possesses low actual speedups ($\leq 1.3$) and requires fixed sparse ratios, meaning that other ratios, such as 4:8, 8:16, or those exceeding 50% sparsity, do not incur any speedups on GPUs. Recent studies suggest that V:N:M sparsity is promising in addressing these limitations of 2:4 sparsity. However, regarding accuracy, the effects of V:N:M sparsity on broader Transformer models, such as vision Transformers and large language models (LLMs), are largely unexamined. Moreover, Some specific issues related to V:N:M sparsity, such as how to select appropriate V and M values, remain unresolved. In this study, we thoroughly investigate the application of V:N:M sparsity in vision models and LLMs across multiple tasks, from pertaining to downstream tasks. We propose three key approaches to enhance the applicability and accuracy of V:N:M-sparse Transformers, including heuristic V and M selection, V:N:M-specific channel permutation, and three-staged LoRA training techniques. Experimental results show that, with our methods, the DeiT-small achieves lossless accuracy at 64:2:5 sparsity, while the DeiT-base maintains accuracy even at 64:2:8 sparsity. In addition, the fine-tuned LLama2-7B at 64:2:5 sparsity performs comparably or better than training-free 2:4 sparse alternatives on downstream tasks. More importantly, V:N:M-sparse Transformers offer a wider range of speedup-accuracy trade-offs compared to 2:4 sparsity. Overall, our exploration largely facilitates the V:N:M sparsity to act as a truly effective acceleration solution for Transformers in cost-sensitive inference scenarios.

☆ SMART: Self-learning Meta-strategy Agent for Reasoning Tasks

Rongxing Liu, Kumar Shridhar, Manish Prajapat, Patrick Xia, Mrinmaya Sachan

Tasks requiring deductive reasoning, especially those involving multiple steps, often demand adaptive strategies such as intermediate generation of rationales or programs, as no single approach is universally optimal. While Language Models (LMs) can enhance their outputs through iterative self-refinement and strategy adjustments, they frequently fail to apply the most effective strategy in their first attempt. This inefficiency raises the question: Can LMs learn to select the optimal strategy in the first attempt, without a need for refinement? To address this challenge, we introduce SMART (Self-learning Meta-strategy Agent for Reasoning Tasks), a novel framework that enables LMs to autonomously learn and select the most effective strategies for various reasoning tasks. We model the strategy selection process as a Markov Decision Process and leverage reinforcement learning-driven continuous self-improvement to allow the model to find the suitable strategy to solve a given task. Unlike traditional self-refinement methods that rely on multiple inference passes or external feedback, SMART allows an LM to internalize the outcomes of its own reasoning processes and adjust its strategy accordingly, aiming for correct solutions on the first attempt. Our experiments across various reasoning datasets and with different model architectures demonstrate that SMART significantly enhances the ability of models to choose optimal strategies without external guidance (+15 points on the GSM8K dataset). By achieving higher accuracy with a single inference pass, SMART not only improves performance but also reduces computational costs for refinement-based strategies, paving the way for more efficient and intelligent reasoning in LMs.

☆ MNIST-Nd: a set of naturalistic datasets to benchmark clustering across dimensions

Polina Turishcheva, Laura Hansel, Martin Ritzert, Marissa A. Weis, Alexander S. Ecker

Driven by advances in recording technology, large-scale high-dimensional datasets have emerged across many scientific disciplines. Especially in biology, clustering is often used to gain insights into the structure of such datasets, for instance to understand the organization of different cell types. However, clustering is known to scale poorly to high dimensions, even though the exact impact of dimensionality is unclear as current benchmark datasets are mostly two-dimensional. Here we propose MNIST-Nd, a set of synthetic datasets that share a key property of real-world datasets, namely that individual samples are noisy and clusters do not perfectly separate. MNIST-Nd is obtained by training mixture variational autoencoders with 2 to 64 latent dimensions on MNIST, resulting in six datasets with comparable structure but varying dimensionality. It thus offers the chance to disentangle the impact of dimensionality on clustering. Preliminary common clustering algorithm benchmarks on MNIST-Nd suggest that Leiden is the most robust for growing dimensions.

☆ Integer linear programming for unsupervised training set selection in molecular machine learning

Matthieu Haeberle, Puck van Gerwen, Ruben Laplaza, Ksenia R. Briling, Jan Weinreich, Friedrich Eisenbrand, Clemence Corminboeuf

Integer linear programming (ILP) is an elegant approach to solve linear optimization problems, naturally described using integer decision variables. Within the context of physics-inspired machine learning applied to chemistry, we demonstrate the relevance of an ILP formulation to select molecular training sets for predictions of size-extensive properties. We show that our algorithm outperforms existing unsupervised training set selection approaches, especially when predicting properties of molecules larger than those present in the training set. We argue that the reason for the improved performance is due to the selection that is based on the notion of local similarity (i.e., per-atom) and a unique ILP approach that finds optimal solutions efficiently. Altogether, this work provides a practical algorithm to improve the performance of physics-inspired machine learning models and offers insights into the conceptual differences with existing training set selection approaches.

comment: 31 pages + SI (15 pages)

☆ Extracting Spatiotemporal Data from Gradients with Large Language Models

Lele Zheng, Yang Cao, Renhe Jiang, Kenjiro Taura, Yulong Shen, Sheng Li, Masatoshi Yoshikawa

Recent works show that sensitive user data can be reconstructed from gradient updates, breaking the key privacy promise of federated learning. While success was demonstrated primarily on image data, these methods do not directly transfer to other domains, such as spatiotemporal data. To understand privacy risks in spatiotemporal federated learning, we first propose Spatiotemporal Gradient Inversion Attack (ST-GIA), a gradient attack algorithm tailored to spatiotemporal data that successfully reconstructs the original location from gradients. Furthermore, the absence of priors in attacks on spatiotemporal data has hindered the accurate reconstruction of real client data. To address this limitation, we propose ST-GIA+, which utilizes an auxiliary language model to guide the search for potential locations, thereby successfully reconstructing the original data from gradients. In addition, we design an adaptive defense strategy to mitigate gradient inversion attacks in spatiotemporal federated learning. By dynamically adjusting the perturbation levels, we can offer tailored protection for varying rounds of training data, thereby achieving a better trade-off between privacy and utility than current state-of-the-art methods. Through intensive experimental analysis on three real-world datasets, we reveal that the proposed defense strategy can well preserve the utility of spatiotemporal federated learning with effective security protection.

comment: arXiv admin note: substantial text overlap with arXiv:2407.08529

☆ SeaDAG: Semi-autoregressive Diffusion for Conditional Directed Acyclic Graph Generation

Xinyi Zhou, Xing Li, Yingzhao Lian, Yiwen Wang, Lei Chen, Mingxuan Yuan, Jianye Hao, Guangyong Chen, Pheng Ann Heng

We introduce SeaDAG, a semi-autoregressive diffusion model for conditional generation of Directed Acyclic Graphs (DAGs). Considering their inherent layer-wise structure, we simulate layer-wise autoregressive generation by designing different denoising speed for different layers. Unlike conventional autoregressive generation that lacks a global graph structure view, our method maintains a complete graph structure at each diffusion step, enabling operations such as property control that require the full graph structure. Leveraging this capability, we evaluate the DAG properties during training by employing a graph property decoder. We explicitly train the model to learn graph conditioning with a condition loss, which enhances the diffusion model's capacity to generate graphs that are both realistic and aligned with specified properties. We evaluate our method on two representative conditional DAG generation tasks: (1) circuit generation from truth tables, where precise DAG structures are crucial for realizing circuit functionality, and (2) molecule generation based on quantum properties. Our approach demonstrates promising results, generating high-quality and realistic DAGs that closely align with given conditions.

☆ Statistical Inference for Temporal Difference Learning with Linear Function Approximation

Weichen Wu, Gen Li, Yuting Wei, Alessandro Rinaldo

Statistical inference with finite-sample validity for the value function of a given policy in Markov decision processes (MDPs) is crucial for ensuring the reliability of reinforcement learning. Temporal Difference (TD) learning, arguably the most widely used algorithm for policy evaluation, serves as a natural framework for this purpose.In this paper, we study the consistency properties of TD learning with Polyak-Ruppert averaging and linear function approximation, and obtain three significant improvements over existing results. First, we derive a novel sharp high-dimensional probability convergence guarantee that depends explicitly on the asymptotic variance and holds under weak conditions. We further establish refined high-dimensional Berry-Esseen bounds over the class of convex sets that guarantee faster rates than those in the literature. Finally, we propose a plug-in estimator for the asymptotic covariance matrix, designed for efficient online computation. These results enable the construction of confidence regions and simultaneous confidence intervals for the linear parameters of the value function, with guaranteed finite-sample coverage. We demonstrate the applicability of our theoretical findings through numerical experiments.

☆ Addressing Spectral Bias of Deep Neural Networks by Multi-Grade Deep Learning

Ronglong Fang, Yuesheng Xu

Deep neural networks (DNNs) suffer from the spectral bias, wherein DNNs typically exhibit a tendency to prioritize the learning of lower-frequency components of a function, struggling to capture its high-frequency features. This paper is to address this issue. Notice that a function having only low frequency components may be well-represented by a shallow neural network (SNN), a network having only a few layers. By observing that composition of low frequency functions can effectively approximate a high-frequency function, we propose to learn a function containing high-frequency components by composing several SNNs, each of which learns certain low-frequency information from the given data. We implement the proposed idea by exploiting the multi-grade deep learning (MGDL) model, a recently introduced model that trains a DNN incrementally, grade by grade, a current grade learning from the residue of the previous grade only an SNN composed with the SNNs trained in the preceding grades as features. We apply MGDL to synthetic, manifold, colored images, and MNIST datasets, all characterized by presence of high-frequency features. Our study reveals that MGDL excels at representing functions containing high-frequency information. Specifically, the neural networks learned in each grade adeptly capture some low-frequency information, allowing their compositions with SNNs learned in the previous grades effectively representing the high-frequency features. Our experimental results underscore the efficacy of MGDL in addressing the spectral bias inherent in DNNs. By leveraging MGDL, we offer insights into overcoming spectral bias limitation of DNNs, thereby enhancing the performance and applicability of deep learning models in tasks requiring the representation of high-frequency information. This study confirms that the proposed method offers a promising solution to address the spectral bias of DNNs.

☆ LDAdam: Adaptive Optimization from Low-Dimensional Gradient Statistics

Thomas Robert, Mher Safaryan, Ionut-Vlad Modoranu, Dan Alistarh

We introduce LDAdam, a memory-efficient optimizer for training large models, that performs adaptive optimization steps within lower dimensional subspaces, while consistently exploring the full parameter space during training. This strategy keeps the optimizer's memory footprint to a fraction of the model size. LDAdam relies on a new projection-aware update rule for the optimizer states that allows for transitioning between subspaces, i.e., estimation of the statistics of the projected gradients. To mitigate the errors due to low-rank projection, LDAdam integrates a new generalized error feedback mechanism, which explicitly accounts for both gradient and optimizer state compression. We prove the convergence of LDAdam under standard assumptions, and show that LDAdam allows for accurate and efficient fine-tuning and pre-training of language models.

comment: 36 pages

☆ ExDBN: Exact learning of Dynamic Bayesian Networks

Pavel Rytíř, Aleš Wodecki, Georgios Korpas, Jakub Mareček

Causal learning from data has received much attention in recent years. One way of capturing causal relationships is by utilizing Bayesian networks. There, one recovers a weighted directed acyclic graph, in which random variables are represented by vertices, and the weights associated with each edge represent the strengths of the causal relationships between them. This concept is extended to capture dynamic effects by introducing a dependency on past data, which may be captured by the structural equation model, which is utilized in the present contribution to formulate a score-based learning approach. A mixed-integer quadratic program is formulated and an algorithmic solution proposed, in which the pre-generation of exponentially many acyclicity constraints is avoided by utilizing the so-called branch-and-cut ("lazy constraint") method. Comparing the novel approach to the state of the art, we show that the proposed approach turns out to produce excellent results when applied to small and medium-sized synthetic instances of up to 25 time-series. Lastly, two interesting applications in bio-science and finance, to which the method is directly applied, further stress the opportunities in developing highly accurate, globally convergent solvers that can handle modest instances.

comment: 12 pages

☆ CartesianMoE: Boosting Knowledge Sharing among Experts via Cartesian Product Routing in Mixture-of-Experts

Zhenpeng Su, Xing Wu, Zijia Lin, Yizhe Xiong, Minxuan Lv, Guangyuan Ma, Hui Chen, Songlin Hu, Guiguang Ding

Large language models (LLM) have been attracting much attention from the community recently, due to their remarkable performance in all kinds of downstream tasks. According to the well-known scaling law, scaling up a dense LLM enhances its capabilities, but also significantly increases the computational complexity. Mixture-of-Experts (MoE) models address that by allowing the model size to grow without substantially raising training or inference costs. Yet MoE models face challenges regarding knowledge sharing among experts, making their performance somehow sensitive to routing accuracy. To tackle that, previous works introduced shared experts and combined their outputs with those of the top $K$ routed experts in an ``addition'' manner. In this paper, inspired by collective matrix factorization to learn shared knowledge among data, we propose CartesianMoE, which implements more effective knowledge sharing among experts in more like a ``multiplication'' manner. Extensive experimental results indicate that CartesianMoE outperforms previous MoE models for building LLMs, in terms of both perplexity and downstream task performance. And we also find that CartesianMoE achieves better expert routing robustness.

☆ On the Geometry of Regularization in Adversarial Training: High-Dimensional Asymptotics and Generalization Bounds

Matteo Vilucchio, Nikolaos Tsilivis, Bruno Loureiro, Julia Kempe

Regularization, whether explicit in terms of a penalty in the loss or implicit in the choice of algorithm, is a cornerstone of modern machine learning. Indeed, controlling the complexity of the model class is particularly important when data is scarce, noisy or contaminated, as it translates a statistical belief on the underlying structure of the data. This work investigates the question of how to choose the regularization norm $\lVert \cdot \rVert$ in the context of high-dimensional adversarial training for binary classification. To this end, we first derive an exact asymptotic description of the robust, regularized empirical risk minimizer for various types of adversarial attacks and regularization norms (including non-$\ell_p$ norms). We complement this analysis with a uniform convergence analysis, deriving bounds on the Rademacher Complexity for this class of problems. Leveraging our theoretical results, we quantitatively characterize the relationship between perturbation size and the optimal choice of $\lVert \cdot \rVert$, confirming the intuition that, in the data scarce regime, the type of regularization becomes increasingly important for adversarial training as perturbations grow in size.

☆ Near-Optimal Algorithm for Non-Stationary Kernelized Bandits

Shogo Iwazaki, Shion Takeno

This paper studies a non-stationary kernelized bandit (KB) problem, also called time-varying Bayesian optimization, where one seeks to minimize the regret under an unknown reward function that varies over time. In particular, we focus on a near-optimal algorithm whose regret upper bound matches the regret lower bound. For this goal, we show the first algorithm-independent regret lower bound for non-stationary KB with squared exponential and Mat\'ern kernels, which reveals that an existing optimization-based KB algorithm with slight modification is near-optimal. However, this existing algorithm suffers from feasibility issues due to its huge computational cost. Therefore, we propose a novel near-optimal algorithm called restarting phased elimination with random permutation (R-PERP), which bypasses the huge computational cost. A technical key point is the simple permutation procedures of query candidates, which enable us to derive a novel tighter confidence bound tailored to the non-stationary problems.

comment: 24 pages, 2 figures

☆ GFlowNets for Hamiltonian decomposition in groups of compatible operators NeurIPS 2024

Isaac L. Huidobro-Meezs, Jun Dai, Guillaume Rabusseau, Rodrigo A. Vargas-Hernández

Quantum computing presents a promising alternative for the direct simulation of quantum systems with the potential to explore chemical problems beyond the capabilities of classical methods. However, current quantum algorithms are constrained by hardware limitations and the increased number of measurements required to achieve chemical accuracy. To address the measurement challenge, techniques for grouping commuting and anti-commuting terms, driven by heuristics, have been developed to reduce the number of measurements needed in quantum algorithms on near-term quantum devices. In this work, we propose a probabilistic framework using GFlowNets to group fully (FC) or qubit-wise commuting (QWC) terms within a given Hamiltonian. The significance of this approach is demonstrated by the reduced number of measurements for the found groupings; 51% and 67% reduction factors respectively for FC and QWC partitionings with respect to greedy coloring algorithms, highlighting the potential of GFlowNets for future applications in the measurement problem. Furthermore, the flexibility of our algorithm extends its applicability to other resource optimization problems in Hamiltonian simulation, such as circuit design.

comment: 8 pages, 2 figures. Accepted for Machine Learning and the Physical Sciences Workshop, NeurIPS 2024. Submission Number: 167

☆ TimeMixer++: A General Time Series Pattern Machine for Universal Predictive Analysis

Shiyu Wang, Jiawei Li, Xiaoming Shi, Zhou Ye, Baichuan Mo, Wenze Lin, Shengtong Ju, Zhixuan Chu, Ming Jin

Time series analysis plays a critical role in numerous applications, supporting tasks such as forecasting, classification, anomaly detection, and imputation. In this work, we present the time series pattern machine (TSPM), a model designed to excel in a broad range of time series tasks through powerful representation and pattern extraction capabilities. Traditional time series models often struggle to capture universal patterns, limiting their effectiveness across diverse tasks. To address this, we define multiple scales in the time domain and various resolutions in the frequency domain, employing various mixing strategies to extract intricate, task-adaptive time series patterns. Specifically, we introduce a general-purpose TSPM that processes multi-scale time series using (1) multi-resolution time imaging (MRTI), (2) time image decomposition (TID), (3) multi-scale mixing (MCM), and (4) multi-resolution mixing (MRM) to extract comprehensive temporal patterns. MRTI transforms multi-scale time series into multi-resolution time images, capturing patterns across both temporal and frequency domains. TID leverages dual-axis attention to extract seasonal and trend patterns, while MCM hierarchically aggregates these patterns across scales. MRM adaptively integrates all representations across resolutions. This method achieves state-of-the-art performance across 8 time series analytical tasks, consistently surpassing both general-purpose and task-specific models. Our work marks a promising step toward the next generation of TSPMs, paving the way for further advancements in time series analysis.

☆ Natural GaLore: Accelerating GaLore for memory-efficient LLM Training and Fine-tuning

Arijit Das

Training LLMs presents significant memory challenges due to growing size of data, weights, and optimizer states. Techniques such as data and model parallelism, gradient checkpointing, and offloading strategies address this issue but are often infeasible due to hardware constraints. To mitigate memory usage, alternative methods like Parameter-Efficient-Fine-Tuning (PEFT) and GaLore approximate weights or optimizer states. PEFT methods, such as LoRA, have gained popularity for fine-tuning LLMs, though they require a full-rank warm start. In contrast, GaLore allows full-parameter learning while being more memory-efficient. This work introduces Natural GaLore, a simple drop in replacement for AdamW, which efficiently applies the inverse Empirical Fisher Information Matrix to low-rank gradients using Woodbury's Identity. We demonstrate that incorporating second-order information speeds up optimization significantly, especially when the iteration budget is limited. Empirical pretraining on 60M, 130M, 350M, and 1.1B parameter Llama models on C4 data demonstrate significantly lower perplexity over GaLore without additional memory overhead. By fine-tuning RoBERTa on the GLUE benchmark using Natural GaLore, we demonstrate significant reduction in gap 86.05% vs 86.28% for full-finetuning. Furthermore, fine-tuning the TinyLlama 1.1B model for function calling using the TinyAgent framework shows that Natural GaLore achieving 83.09% accuracy on the TinyAgent dataset, significantly outperforms 16-bit LoRA at 80.06% and even surpasses GPT4-Turbo by 4%, all while using 30% less memory. All code to reproduce the results are available at: https://github.com/selfsupervised-ai/Natural-GaLore.git

comment: 10 pages, 3 tables, 3 figures

☆ Information-Theoretic Minimax Regret Bounds for Reinforcement Learning based on Duality

Raghav Bongole, Amaury Gouverneur, Borja Rodríguez-Gálvez, Tobias J. Oechtering, Mikael Skoglund

We study agents acting in an unknown environment where the agent's goal is to find a robust policy. We consider robust policies as policies that achieve high cumulative rewards for all possible environments. To this end, we consider agents minimizing the maximum regret over different environment parameters, leading to the study of minimax regret. This research focuses on deriving information-theoretic bounds for minimax regret in Markov Decision Processes (MDPs) with a finite time horizon. Building on concepts from supervised learning, such as minimum excess risk (MER) and minimax excess risk, we use recent bounds on the Bayesian regret to derive minimax regret bounds. Specifically, we establish minimax theorems and use bounds on the Bayesian regret to perform minimax regret analysis using these minimax theorems. Our contributions include defining a suitable minimax regret in the context of MDPs, finding information-theoretic bounds for it, and applying these bounds in various scenarios.

☆ Massimo: Public Queue Monitoring and Management using Mass-Spring Model

Abhijeet Kumar, Unnati Singh, Rajdeep Chatterjee, Tathagata Bandyopadhyay

An efficient system of a queue control and regulation in public spaces is very important in order to avoid the traffic jams and to improve the customer satisfaction. This article offers a detailed road map based on a merger of intelligent systems and creating an efficient systems of queues in public places. Through the utilization of different technologies i.e. computer vision, machine learning algorithms, deep learning our system provide accurate information about the place is crowded or not and the necessary efforts to be taken.

comment: 8 pages, 6 figures, 3 algorithms, 3 tables

☆ Resilient Temporal GCN for Smart Grid State Estimation Under Topology Inaccuracies

Seyed Hamed Haghshenas, Mia Naeini

State Estimation is a crucial task in power systems. Graph Neural Networks have demonstrated significant potential in state estimation for power systems by effectively analyzing measurement data and capturing the complex interactions and interrelations among the measurements through the system's graph structure. However, the information about the system's graph structure may be inaccurate due to noise, attack or lack of accurate information about the topology of the system. This paper studies these scenarios under topology uncertainties and evaluates the impact of the topology uncertainties on the performance of a Temporal Graph Convolutional Network (TGCN) for state estimation in power systems. In order to make the model resilient to topology uncertainties, modifications in the TGCN model are proposed to incorporate a knowledge graph, generated based on the measurement data. This knowledge graph supports the assumed uncertain system graph. Two variations of the TGCN architecture are introduced to integrate the knowledge graph, and their performances are evaluated and compared to demonstrate improved resilience against topology uncertainties. The evaluation results indicate that while the two proposed architecture show different performance, they both improve the performance of the TGCN state estimation under topology uncertainties.

comment: 9 pages, 5 figures

☆ 1024m at SMM4H 2024: Tasks 3, 5 & 6 -- Ensembles of Transformers and Large Language Models for Medical Text Classification

Ram Mohan Rao Kadiyala, M. V. P. Chandra Sekhara Rao

Social media is a great source of data for users reporting information and regarding their health and how various things have had an effect on them. This paper presents various approaches using Transformers and Large Language Models and their ensembles, their performance along with advantages and drawbacks for various tasks of SMM4H'24 - Classifying texts on impact of nature and outdoor spaces on the author's mental health (Task 3), Binary classification of tweets reporting their children's health disorders like Asthma, Autism, ADHD and Speech disorder (task 5), Binary classification of users self-reporting their age (task 6).

comment: short paper , acl 2024

☆ MultiRC: Joint Learning for Time Series Anomaly Prediction and Detection with Multi-scale Reconstructive Contrast

Shiyan Hu, Kai Zhao, Xiangfei Qiu, Yang Shu, Jilin Hu, Bin Yang, Chenjuan Guo

Many methods have been proposed for unsupervised time series anomaly detection. Despite some progress, research on predicting future anomalies is still relatively scarce. Predicting anomalies is particularly challenging due to the diverse reaction time and the lack of labeled data. To address these challenges, we propose MultiRC to integrate reconstructive and contrastive learning for joint learning of anomaly prediction and detection, with multi-scale structure and adaptive dominant period mask to deal with the diverse reaction time. MultiRC also generates negative samples to provide essential training momentum for the anomaly prediction tasks and prevent model degradation. We evaluate seven benchmark datasets from different fields. For both anomaly prediction and detection tasks, MultiRC outperforms existing state-of-the-art methods.

☆ Augmenting Legal Decision Support Systems with LLM-based NLI for Analyzing Social Media Evidence

Ram Mohan Rao Kadiyala, Siddartha Pullakhandam, Kanwal Mehreen, Subhasya Tippareddy, Ashay Srivastava

This paper presents our system description and error analysis of our entry for NLLP 2024 shared task on Legal Natural Language Inference (L-NLI) \citep{hagag2024legallenssharedtask2024}. The task required classifying these relationships as entailed, contradicted, or neutral, indicating any association between the review and the complaint. Our system emerged as the winning submission, significantly outperforming other entries with a substantial margin and demonstrating the effectiveness of our approach in legal text analysis. We provide a detailed analysis of the strengths and limitations of each model and approach tested, along with a thorough error analysis and suggestions for future improvements. This paper aims to contribute to the growing field of legal NLP by offering insights into advanced techniques for natural language inference in legal contexts, making it accessible to both experts and newcomers in the field.

comment: 8 pages , accepted to emnlp 2024

☆ Analyzing Closed-loop Training Techniques for Realistic Traffic Agent Models in Autonomous Highway Driving Simulations

Matthias Bitzer, Reinis Cimurs, Benjamin Coors, Johannes Goth, Sebastian Ziesche, Philipp Geiger, Maximilian Naumann

Simulation plays a crucial role in the rapid development and safe deployment of autonomous vehicles. Realistic traffic agent models are indispensable for bridging the gap between simulation and the real world. Many existing approaches for imitating human behavior are based on learning from demonstration. However, these approaches are often constrained by focusing on individual training strategies. Therefore, to foster a broader understanding of realistic traffic agent modeling, in this paper, we provide an extensive comparative analysis of different training principles, with a focus on closed-loop methods for highway driving simulation. We experimentally compare (i) open-loop vs. closed-loop multi-agent training, (ii) adversarial vs. deterministic supervised training, (iii) the impact of reinforcement losses, and (iv) the impact of training alongside log-replayed agents to identify suitable training techniques for realistic agent modeling. Furthermore, we identify promising combinations of different closed-loop training methods.

comment: 15 pages, 6 figures, 4 tables

☆ A quantitative Robbins-Siegmund theorem

Morenikeji Neri, Thomas Powell

The Robbins-Siegmund theorem is one of the most important results in stochastic optimization, where it is widely used to prove the convergence of stochastic algorithms. We provide a quantitative version of the theorem, establishing a bound on how far one needs to look in order to locate a region of metastability in the sense of Tao. Our proof involves a metastable analogue of Doob's theorem for $L_1$-supermartingales along with a series of technical lemmas that make precise how quantitative information propagates through sums and products of stochastic processes. In this way, our paper establishes a general methodology for finding metastable bounds for stochastic processes that can be reduced to supermartingales, and therefore for obtaining quantitative convergence information across a broad class of stochastic algorithms whose convergence proof relies on some variation of the Robbins-Siegmund theorem. We conclude by discussing how our general quantitative result might be used in practice.

comment: 30 pages

☆ State Estimation Using Sparse DEIM and Recurrent Neural Networks

Mohammad Farazmand

Discrete Empirical Interpolation Method (DEIM) estimates a function from its pointwise incomplete observations. In particular, this method can be used to estimate the state of a dynamical system from observational data gathered by sensors. However, when the number of observations are limited, DEIM returns large estimation errors. Sparse DEIM (S-DEIM) was recently developed to address this problem by introducing a kernel vector which previous DEIM-based methods had ignored. Unfortunately, estimating the optimal kernel vector in S-DEIM is a difficult task. Here, we introduce a data-driven method to estimate this kernel vector from sparse observational time series using recurrent neural networks. Using numerical examples, we demonstrate that this machine learning approach together with S-DEIM leads to nearly optimal state estimations.

☆ Visual Representation Learning Guided By Multi-modal Prior Knowledge

Hongkuan Zhou, Lavdim Halilaj, Sebastian Monka, Stefan Schmid, Yuqicheng Zhu, Bo Xiong, Steffen Staab

Despite the remarkable success of deep neural networks (DNNs) in computer vision, they fail to remain high-performing when facing distribution shifts between training and testing data. In this paper, we propose Knowledge-Guided Visual representation learning (KGV), a distribution-based learning approach leveraging multi-modal prior knowledge, to improve generalization under distribution shift. We use prior knowledge from two distinct modalities: 1) a knowledge graph (KG) with hierarchical and association relationships; and 2) generated synthetic images of visual elements semantically represented in the KG. The respective embeddings are generated from the given modalities in a common latent space, i.e., visual embeddings from original and synthetic images as well as knowledge graph embeddings (KGEs). These embeddings are aligned via a novel variant of translation-based KGE methods, where the node and relation embeddings of the KG are modeled as Gaussian distributions and translations respectively. We claim that incorporating multi-model prior knowledge enables more regularized learning of image representations. Thus, the models are able to better generalize across different data distributions. We evaluate KGV on different image classification tasks with major or minor distribution shifts, namely road sign classification across datasets from Germany, China, and Russia, image classification with the mini-ImageNet dataset and its variants, as well as the DVM-CAR dataset. The results demonstrate that KGV consistently exhibits higher accuracy and data efficiency than the baselines across all experiments.

☆ Large Language Models for Cross-lingual Emotion Detection

Ram Mohan Rao Kadiyala

This paper presents a detailed system description of our entry for the WASSA 2024 Task 2, focused on cross-lingual emotion detection. We utilized a combination of large language models (LLMs) and their ensembles to effectively understand and categorize emotions across different languages. Our approach not only outperformed other submissions with a large margin, but also demonstrated the strength of integrating multiple models to enhance performance. Additionally, We conducted a thorough comparison of the benefits and limitations of each model used. An error analysis is included along with suggested areas for future improvement. This paper aims to offer a clear and comprehensive understanding of advanced techniques in emotion detection, making it accessible even to those new to the field.

comment: 6 pages , accepted to acl 2024

☆ Karush-Kuhn-Tucker Condition-Trained Neural Networks (KKT Nets)

Shreya Arvind, Rishabh Pomaje, Rajshekhar V Bhat

This paper presents a novel approach to solving convex optimization problems by leveraging the fact that, under certain regularity conditions, any set of primal or dual variables satisfying the Karush-Kuhn-Tucker (KKT) conditions is necessary and sufficient for optimality. Similar to Theory-Trained Neural Networks (TTNNs), the parameters of the convex optimization problem are input to the neural network, and the expected outputs are the optimal primal and dual variables. A choice for the loss function in this case is a loss, which we refer to as the KKT Loss, that measures how well the network's outputs satisfy the KKT conditions. We demonstrate the effectiveness of this approach using a linear program as an example. For this problem, we observe that minimizing the KKT Loss alone outperforms training the network with a weighted sum of the KKT Loss and a Data Loss (the mean-squared error between the ground truth optimal solutions and the network's output). Moreover, minimizing only the Data Loss yields inferior results compared to those obtained by minimizing the KKT Loss. While the approach is promising, the obtained primal and dual solutions are not sufficiently close to the ground truth optimal solutions. In the future, we aim to develop improved models to obtain solutions closer to the ground truth and extend the approach to other problem classes.

☆ TS-ACL: A Time Series Analytic Continual Learning Framework for Privacy-Preserving and Class-Incremental Pattern Recognition

Kejia Fan, Jiaxu Li, Songning Lai, Linpu Lv, Anfeng Liu, Jianheng Tang, Houbing Herbert Song, Huiping Zhuang

Class-incremental Learning (CIL) in Time Series Classification (TSC) aims to incrementally train models using the streaming time series data that arrives continuously. The main problem in this scenario is catastrophic forgetting, i.e., training models with new samples inevitably leads to the forgetting of previously learned knowledge. Among existing methods, the replay-based methods achieve satisfactory performance but compromise privacy, while exemplar-free methods protect privacy but suffer from low accuracy. However, more critically, owing to their reliance on gradient-based update techniques, these existing methods fundamentally cannot solve the catastrophic forgetting problem. In TSC scenarios with continuously arriving data and temporally shifting distributions, these methods become even less practical. In this paper, we propose a Time Series Analytic Continual Learning framework, called TS-ACL. Inspired by analytical learning, TS-ACL transforms neural network updates into gradient-free linear regression problems, thereby fundamentally mitigating catastrophic forgetting. Specifically, employing a pre-trained and frozen feature extraction encoder, TS-ACL only needs to update its analytic classifier recursively in a lightweight manner that is highly suitable for real-time applications and large-scale data processing. Additionally, we theoretically demonstrate that the model obtained recursively through the TS-ACL is exactly equivalent to a model trained on the complete dataset in a centralized manner, thereby establishing the property of absolute knowledge memory. Extensive experiments validate the superior performance of our TS-ACL.

comment: 11 pages, 3 figures, 2 tables

☆ User-centric evaluation of explainability of AI with and for humans: a comprehensive empirical study

Szymon Bobek, Paloma Korycińska, Monika Krakowska, Maciej Mozolewski, Dorota Rak, Magdalena Zych, Magdalena Wójcik, Grzegorz J. Nalepa

This study is located in the Human-Centered Artificial Intelligence (HCAI) and focuses on the results of a user-centered assessment of commonly used eXplainable Artificial Intelligence (XAI) algorithms, specifically investigating how humans understand and interact with the explanations provided by these algorithms. To achieve this, we employed a multi-disciplinary approach that included state-of-the-art research methods from social sciences to measure the comprehensibility of explanations generated by a state-of-the-art lachine learning model, specifically the Gradient Boosting Classifier (XGBClassifier). We conducted an extensive empirical user study involving interviews with 39 participants from three different groups, each with varying expertise in data science, data visualization, and domain-specific knowledge related to the dataset used for training the machine learning model. Participants were asked a series of questions to assess their understanding of the model's explanations. To ensure replicability, we built the model using a publicly available dataset from the UC Irvine Machine Learning Repository, focusing on edible and non-edible mushrooms. Our findings reveal limitations in existing XAI methods and confirm the need for new design principles and evaluation techniques that address the specific information needs and user perspectives of different classes of AI stakeholders. We believe that the results of our research and the cross-disciplinary methodology we developed can be successfully adapted to various data types and user profiles, thus promoting dialogue and address opportunities in HCAI research. To support this, we are making the data resulting from our study publicly available.

☆ GReFEL: Geometry-Aware Reliable Facial Expression Learning under Bias and Imbalanced Data Distribution ACCV 2024

Azmine Toushik Wasi, Taki Hasan Rafi, Raima Islam, Karlo Serbetar, Dong Kyu Chae

Reliable facial expression learning (FEL) involves the effective learning of distinctive facial expression characteristics for more reliable, unbiased and accurate predictions in real-life settings. However, current systems struggle with FEL tasks because of the variance in people's facial expressions due to their unique facial structures, movements, tones, and demographics. Biased and imbalanced datasets compound this challenge, leading to wrong and biased prediction labels. To tackle these, we introduce GReFEL, leveraging Vision Transformers and a facial geometry-aware anchor-based reliability balancing module to combat imbalanced data distributions, bias, and uncertainty in facial expression learning. Integrating local and global data with anchors that learn different facial data points and structural features, our approach adjusts biased and mislabeled emotions caused by intra-class disparity, inter-class similarity, and scale sensitivity, resulting in comprehensive, accurate, and reliable facial expression predictions. Our model outperforms current state-of-the-art methodologies, as demonstrated by extensive experiments on various datasets.

comment: ACCV 2024. Extended version of ARBEx (arXiv:2305.01486)

☆ Automatic Differentiation of Optimization Algorithms with Time-Varying Updates

Sheheryar Mehmood, Peter Ochs

Numerous Optimization Algorithms have a time-varying update rule thanks to, for instance, a changing step size, momentum parameter or, Hessian approximation. In this paper, we apply unrolled or automatic differentiation to a time-varying iterative process and provide convergence (rate) guarantees for the resulting derivative iterates. We adapt these convergence results and apply them to proximal gradient descent with variable step size and FISTA when solving partly smooth problems. We confirm our findings numerically by solving $\ell_1$ and $\ell_2$-regularized linear and logisitc regression respectively. Our theoretical and numerical results show that the convergence rate of the algorithm is reflected in its derivative iterates.

★ Diverse Policies Recovering via Pointwise Mutual Information Weighted Imitation Learning

Hanlin Yang, Jian Yao, Weiming Liu, Qing Wang, Hanmin Qin, Hansheng Kong, Kirk Tang, Jiechao Xiong, Chao Yu, Kai Li, Junliang Xing, Hongwu Chen, Juchao Zhuo, Qiang Fu, Yang Wei, Haobo Fu

Recovering a spectrum of diverse policies from a set of expert trajectories is an important research topic in imitation learning. After determining a latent style for a trajectory, previous diverse policies recovering methods usually employ a vanilla behavioral cloning learning objective conditioned on the latent style, treating each state-action pair in the trajectory with equal importance. Based on an observation that in many scenarios, behavioral styles are often highly relevant with only a subset of state-action pairs, this paper presents a new principled method in diverse polices recovery. In particular, after inferring or assigning a latent style for a trajectory, we enhance the vanilla behavioral cloning by incorporating a weighting mechanism based on pointwise mutual information. This additional weighting reflects the significance of each state-action pair's contribution to learning the style, thus allowing our method to focus on state-action pairs most representative of that style. We provide theoretical justifications for our new objective, and extensive empirical evaluations confirm the effectiveness of our method in recovering diverse policies from expert data.

comment: 18 pages, 6 figures

☆ On the Design and Performance of Machine Learning Based Error Correcting Decoders

Yuncheng Yuan, Péter Scheepers, Lydia Tasiou, Yunus Can Gültekin, Federico Corradi, Alex Alvarado

This paper analyzes the design and competitiveness of four neural network (NN) architectures recently proposed as decoders for forward error correction (FEC) codes. We first consider the so-called single-label neural network (SLNN) and the multi-label neural network (MLNN) decoders which have been reported to achieve near maximum likelihood (ML) performance. Here, we show analytically that SLNN and MLNN decoders can always achieve ML performance, regardless of the code dimensions -- although at the cost of computational complexity -- and no training is in fact required. We then turn our attention to two transformer-based decoders: the error correction code transformer (ECCT) and the cross-attention message passing transformer (CrossMPT). We compare their performance against traditional decoders, and show that ordered statistics decoding outperforms these transformer-based decoders. The results in this paper cast serious doubts on the application of NN-based FEC decoders in the short and medium block length regime.

comment: 6 pages, 4 figures, submitted for possible presentation in a conference

☆ Model Mimic Attack: Knowledge Distillation for Provably Transferable Adversarial Examples

Kirill Lukyanov, Andrew Perminov, Denis Turdakov, Mikhail Pautov

The vulnerability of artificial neural networks to adversarial perturbations in the black-box setting is widely studied in the literature. The majority of attack methods to construct these perturbations suffer from an impractically large number of queries required to find an adversarial example. In this work, we focus on knowledge distillation as an approach to conduct transfer-based black-box adversarial attacks and propose an iterative training of the surrogate model on an expanding dataset. This work is the first, to our knowledge, to provide provable guarantees on the success of knowledge distillation-based attack on classification neural networks: we prove that if the student model has enough learning capabilities, the attack on the teacher model is guaranteed to be found within the finite number of distillation iterations.

☆ Using GPT Models for Qualitative and Quantitative News Analytics in the 2024 US Presidental Election Process

Bohdan M. Pavlyshenko

The paper considers an approach of using Google Search API and GPT-4o model for qualitative and quantitative analyses of news through retrieval-augmented generation (RAG). This approach was applied to analyze news about the 2024 US presidential election process. Different news sources for different time periods have been analyzed. Quantitative scores generated by GPT model have been analyzed using Bayesian regression to derive trend lines. The distributions found for the regression parameters allow for the analysis of uncertainty in the election process. The obtained results demonstrate that using the GPT models for news analysis, one can get informative analytics and provide key insights that can be applied in further analyses of election processes.

☆ Distributed Learning for UAV Swarms

Chen Hu, Hanchi Ren, Jingjing Deng, Xianghua Xie

Unmanned Aerial Vehicle (UAV) swarms are increasingly deployed in dynamic, data-rich environments for applications such as environmental monitoring and surveillance. These scenarios demand efficient data processing while maintaining privacy and security, making Federated Learning (FL) a promising solution. FL allows UAVs to collaboratively train global models without sharing raw data, but challenges arise due to the non-Independent and Identically Distributed (non-IID) nature of the data collected by UAVs. In this study, we show an integration of the state-of-the-art FL methods to UAV Swarm application and invetigate the performance of multiple aggregation methods (namely FedAvg, FedProx, FedOpt, and MOON) with a particular focus on tackling non-IID on a variety of datasets, specifically MNIST for baseline performance, CIFAR10 for natural object classification, EuroSAT for environment monitoring, and CelebA for surveillance. These algorithms were selected to cover improved techniques on both client-side updates and global aggregation. Results show that while all algorithms perform comparably on IID data, their performance deteriorates significantly under non-IID conditions. FedProx demonstrated the most stable overall performance, emphasising the importance of regularising local updates in non-IID environments to mitigate drastic deviations in local models.

☆ FlickerFusion: Intra-trajectory Domain Generalizing Multi-Agent RL NeurIPS '24

Woosung Koh, Wonbeen Oh, Siyeol Kim, Suhin Shin, Hyeongjin Kim, Jaein Jang, Junghyun Lee, Se-Young Yun

Multi-agent reinforcement learning has demonstrated significant potential in addressing complex cooperative tasks across various real-world applications. However, existing MARL approaches often rely on the restrictive assumption that the number of entities (e.g., agents, obstacles) remains constant between training and inference. This overlooks scenarios where entities are dynamically removed or added during the inference trajectory -- a common occurrence in real-world environments like search and rescue missions and dynamic combat situations. In this paper, we tackle the challenge of intra-trajectory dynamic entity composition under zero-shot out-of-domain (OOD) generalization, where such dynamic changes cannot be anticipated beforehand. Our empirical studies reveal that existing MARL methods suffer significant performance degradation and increased uncertainty in these scenarios. In response, we propose FlickerFusion, a novel OOD generalization method that acts as a universally applicable augmentation technique for MARL backbone methods. Our results show that FlickerFusion not only achieves superior inference rewards but also uniquely reduces uncertainty vis-\`a-vis the backbone, compared to existing methods. For standardized evaluation, we introduce MPEv2, an enhanced version of Multi Particle Environments (MPE), consisting of 12 benchmarks. Benchmarks, implementations, and trained models are organized and open-sourced at flickerfusion305.github.io, accompanied by ample demo video renderings.

comment: NeurIPS '24 Open-World Agents Workshop

☆ Enabling Asymmetric Knowledge Transfer in Multi-Task Learning with Self-Auxiliaries

Olivier Graffeuille, Yun Sing Koh, Joerg Wicker, Moritz Lehmann

Knowledge transfer in multi-task learning is typically viewed as a dichotomy; positive transfer, which improves the performance of all tasks, or negative transfer, which hinders the performance of all tasks. In this paper, we investigate the understudied problem of asymmetric task relationships, where knowledge transfer aids the learning of certain tasks while hindering the learning of others. We propose an optimisation strategy that includes additional cloned tasks named self-auxiliaries into the learning process to flexibly transfer knowledge between tasks asymmetrically. Our method can exploit asymmetric task relationships, benefiting from the positive transfer component while avoiding the negative transfer component. We demonstrate that asymmetric knowledge transfer provides substantial improvements in performance compared to existing multi-task optimisation strategies on benchmark computer vision problems.

☆ Mesa-Extrapolation: A Weave Position Encoding Method for Enhanced Extrapolation in LLMs NeurIPS 2024

Xin Ma, Yang Liu, Jingjing Liu, Xiaoxu Ma

Large language models (LLMs), although having revolutionized many fields, still suffer from the challenging extrapolation problem, where the inference ability of LLMs sharply declines beyond their max training lengths. In this work, we conduct a theoretical analysis to better understand why No Position Encoding (NoPE) fails outside its effective range, as well as examining the power of Position Encoding (PE) in this context. Our findings reveal that with meticulous weave position, PE can indeed be extended beyond effective range. Our theorems establish that LLMs equipped with weave PE can achieve improved extrapolation performance without additional cost. Furthermore, we introduce a novel weave PE method, Mesa-Extrapolation, which utilizes a chunk-based triangular attention matrix and applies Stair PE to manage the final chunk. This method not only retains competitive performance but also offers substantial benefits such as significantly reduced memory demand and faster inference speed. Extensive experiments validate the effectiveness of Mesa-Extrapolation, demonstrating its potential as a scalable solution to enhancing LLMs applicative reach.

comment: accepted by NeurIPS 2024. arXiv admin note: text overlap with arXiv:2305.19466 by other authors

☆ Towards Optimal Adapter Placement for Efficient Transfer Learning

Aleksandra I. Nowak, Otniel-Bogdan Mercea, Anurag Arnab, Jonas Pfeiffer, Yann Dauphin, Utku Evci

Parameter-efficient transfer learning (PETL) aims to adapt pre-trained models to new downstream tasks while minimizing the number of fine-tuned parameters. Adapters, a popular approach in PETL, inject additional capacity into existing networks by incorporating low-rank projections, achieving performance comparable to full fine-tuning with significantly fewer parameters. This paper investigates the relationship between the placement of an adapter and its performance. We observe that adapter location within a network significantly impacts its effectiveness, and that the optimal placement is task-dependent. To exploit this observation, we introduce an extended search space of adapter connections, including long-range and recurrent adapters. We demonstrate that even randomly selected adapter placements from this expanded space yield improved results, and that high-performing placements often correlate with high gradient rank. Our findings reveal that a small number of strategically placed adapters can match or exceed the performance of the common baseline of adding adapters in every block, opening a new avenue for research into optimal adapter placement strategies.

☆ TEXEL: A neuromorphic processor with on-chip learning for beyond-CMOS device integration

Hugh Greatorex, Ole Richter, Michele Mastella, Madison Cotteret, Philipp Klein, Maxime Fabre, Arianna Rubino, Willian Soares Girão, Junren Chen, Martin Ziegler, Laura Bégon-Lours, Giacomo Indiveri, Elisabetta Chicca

Recent advances in memory technologies, devices and materials have shown great potential for integration into neuromorphic electronic systems. However, a significant gap remains between the development of these materials and the realization of large-scale, fully functional systems. One key challenge is determining which devices and materials are best suited for specific functions and how they can be paired with CMOS circuitry. To address this, we introduce TEXEL, a mixed-signal neuromorphic architecture designed to explore the integration of on-chip learning circuits and novel two- and three-terminal devices. TEXEL serves as an accessible platform to bridge the gap between CMOS-based neuromorphic computation and the latest advancements in emerging devices. In this paper, we demonstrate the readiness of TEXEL for device integration through comprehensive chip measurements and simulations. TEXEL provides a practical system for testing bio-inspired learning algorithms alongside emerging devices, establishing a tangible link between brain-inspired computation and cutting-edge device research.

comment: 17 pages, 7 figures. Supplementary material: 8 pages, 4 figures

☆ R2I-rPPG: A Robust Region of Interest Selection Method for Remote Photoplethysmography to Extract Heart Rate

Sandeep Nagar, Mark Hasegawa-Johnson, David G. Beiser, Narendra Ahuja

The COVID-19 pandemic has underscored the need for low-cost, scalable approaches to measuring contactless vital signs, either during initial triage at a healthcare facility or virtual telemedicine visits. Remote photoplethysmography (rPPG) can accurately estimate heart rate (HR) when applied to close-up videos of healthy volunteers in well-lit laboratory settings. However, results from such highly optimized laboratory studies may not be readily translated to healthcare settings. One significant barrier to the practical application of rPPG in health care is the accurate localization of the region of interest (ROI). Clinical or telemedicine visits may involve sub-optimal lighting, movement artifacts, variable camera angle, and subject distance. This paper presents an rPPG ROI selection method based on 3D facial landmarks and patient head yaw angle. We then demonstrate the robustness of this ROI selection method when coupled to the Plane-Orthogonal-to-Skin (POS) rPPG method when applied to videos of patients presenting to an Emergency Department for respiratory complaints. Our results demonstrate the effectiveness of our proposed approach in improving the accuracy and robustness of rPPG in a challenging clinical environment.

comment: preprint

☆ Focus Where It Matters: Graph Selective State Focused Attention Networks

Shikhar Vashistha, Neetesh Kumar

Traditional graph neural networks (GNNs) lack scalability and lose individual node characteristics due to over-smoothing, especially in the case of deeper networks. This results in sub-optimal feature representation, affecting the model's performance on tasks involving dynamically changing graphs. To address this issue, we present Graph Selective States Focused Attention Networks (GSANs) based neural network architecture for graph-structured data. The GSAN is enabled by multi-head masked self-attention (MHMSA) and selective state space modeling (S3M) layers to overcome the limitations of GNNs. In GSAN, the MHMSA allows GSAN to dynamically emphasize crucial node connections, particularly in evolving graph environments. The S3M layer enables the network to adjust dynamically in changing node states and improving predictions of node behavior in varying contexts without needing primary knowledge of the graph structure. Furthermore, the S3M layer enhances the generalization of unseen structures and interprets how node states influence link importance. With this, GSAN effectively outperforms inductive and transductive tasks and overcomes the issues that traditional GNNs experience. To analyze the performance behavior of GSAN, a set of state-of-the-art comparative experiments are conducted on graphs benchmark datasets, including $Cora$, $Citeseer$, $Pubmed$ network citation, and $protein-protein-interaction$ datasets, as an outcome, GSAN improved the classification accuracy by $1.56\%$, $8.94\%$, $0.37\%$, and $1.54\%$ on $F1-score$ respectively.

☆ Random Token Fusion for Multi-View Medical Diagnosis NeurIPS 2024

Jingyu Guo, Christos Matsoukas, Fredrik Strand, Kevin Smith

In multi-view medical diagnosis, deep learning-based models often fuse information from different imaging perspectives to improve diagnostic performance. However, existing approaches are prone to overfitting and rely heavily on view-specific features, which can lead to trivial solutions. In this work, we introduce Random Token Fusion (RTF), a novel technique designed to enhance multi-view medical image analysis using vision transformers. By integrating randomness into the feature fusion process during training, RTF addresses the issue of overfitting and enhances the robustness and accuracy of diagnostic models without incurring any additional cost at inference. We validate our approach on standard mammography and chest X-ray benchmark datasets. Through extensive experiments, we demonstrate that RTF consistently improves the performance of existing fusion methods, paving the way for a new generation of multi-view medical foundation models.

comment: Originally published at the NeurIPS 2024 Workshop on Advancements In Medical Foundation Models: Explainability, Robustness, Security, and Beyond (AIM-FM)

☆ Modelling Concurrent RTP Flows for End-to-end Predictions of QoS in Real Time Communications

Tailai Song, Paolo Garza, Michela Meo, Maurizio Matteo Munafò

The Real-time Transport Protocol (RTP)-based real-time communications (RTC) applications, exemplified by video conferencing, have experienced an unparalleled surge in popularity and development in recent years. In pursuit of optimizing their performance, the prediction of Quality of Service (QoS) metrics emerges as a pivotal endeavor, bolstering network monitoring and proactive solutions. However, contemporary approaches are confined to individual RTP flows and metrics, falling short in relationship capture and computational efficiency. To this end, we propose Packet-to-Prediction (P2P), a novel deep learning (DL) framework that hinges on raw packets to simultaneously process concurrent RTP flows and perform end-to-end prediction of multiple QoS metrics. Specifically, we implement a streamlined architecture, namely length-free Transformer with cross and neighbourhood attention, capable of handling an unlimited number of RTP flows, and employ a multi-task learning paradigm to forecast four key metrics in a single shot. Our work is based on extensive traffic collected during real video calls, and conclusively, P2P excels comparative models in both prediction performance and temporal efficiency.

☆ Private, Efficient and Scalable Kernel Learning for Medical Image Analysis

Anika Hannemann, Arjhun Swaminathan, Ali Burak Ünal, Mete Akgün

Medical imaging is key in modern medicine. From magnetic resonance imaging (MRI) to microscopic imaging for blood cell detection, diagnostic medical imaging reveals vital insights into patient health. To predict diseases or provide individualized therapies, machine learning techniques like kernel methods have been widely used. Nevertheless, there are multiple challenges for implementing kernel methods. Medical image data often originates from various hospitals and cannot be combined due to privacy concerns, and the high dimensionality of image data presents another significant obstacle. While randomised encoding offers a promising direction, existing methods often struggle with a trade-off between accuracy and efficiency. Addressing the need for efficient privacy-preserving methods on distributed image data, we introduce OKRA (Orthonormal K-fRAmes), a novel randomized encoding-based approach for kernel-based machine learning. This technique, tailored for widely used kernel functions, significantly enhances scalability and speed compared to current state-of-the-art solutions. Through experiments conducted on various clinical image datasets, we evaluated model quality, computational performance, and resource overhead. Additionally, our method outperforms comparable approaches

☆ Explainability of Highly Associated Fuzzy Churn Patterns in Binary Classification PAKDD 2024

D. Y. C. Wang, Lars Arne Jordanger, Jerry Chun-Wei Lin

Customer churn, particularly in the telecommunications sector, influences both costs and profits. As the explainability of models becomes increasingly important, this study emphasizes not only the explainability of customer churn through machine learning models, but also the importance of identifying multivariate patterns and setting soft bounds for intuitive interpretation. The main objective is to use a machine learning model and fuzzy-set theory with top-\textit{k} HUIM to identify highly associated patterns of customer churn with intuitive identification, referred to as Highly Associated Fuzzy Churn Patterns (HAFCP). Moreover, this method aids in uncovering association rules among multiple features across low, medium, and high distributions. Such discoveries are instrumental in enhancing the explainability of findings. Experiments show that when the top-5 HAFCPs are included in five datasets, a mixture of performance results is observed, with some showing notable improvements. It becomes clear that high importance features enhance explanatory power through their distribution and patterns associated with other features. As a result, the study introduces an innovative approach that improves the explainability and effectiveness of customer churn prediction models.

comment: 18 pages single columns, 4 figures, This paper is an extended version of a work originally presented at the 6th International Workshop on Utility-Driven Mining and Learning (held in conjunction with the 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining - PAKDD 2024) on May 7, 2024

☆ LiMTR: Time Series Motion Prediction for Diverse Road Users through Multimodal Feature Integration NeurIPS 2024

Camiel Oerlemans, Bram Grooten, Michiel Braat, Alaa Alassi, Emilia Silvas, Decebal Constantin Mocanu

Predicting the behavior of road users accurately is crucial to enable the safe operation of autonomous vehicles in urban or densely populated areas. Therefore, there has been a growing interest in time series motion prediction research, leading to significant advancements in state-of-the-art techniques in recent years. However, the potential of using LiDAR data to capture more detailed local features, such as a person's gaze or posture, remains largely unexplored. To address this, we develop a novel multimodal approach for motion prediction based on the PointNet foundation model architecture, incorporating local LiDAR features. Evaluation on the Waymo Open Dataset shows a performance improvement of 6.20% and 1.58% in minADE and mAP respectively, when integrated and compared with the previous state-of-the-art MTR. We open-source the code of our LiMTR model.

comment: Accepted at the NeurIPS 2024 workshop Time Series in the Age of Large Models. Code available at https://github.com/Cing2/LiMTR

☆ Solvation Free Energies from Neural Thermodynamic Integration

Bálint Máté, François Fleuret, Tristan Bereau

We propose to compute solvation free energies via thermodynamic integration along a neural-network potential interpolating between two target Hamiltonians. We use a stochastic interpolant to define an interpolation between the distributions at the level of samples and optimize a neural network potential to match the corresponding equilibrium potential at every intermediate time-step. Once the alignment between the interpolating samples and the interpolating potentials is sufficiently accurate, the free-energy difference between the two Hamiltonians can be estimated using (neural) thermodynamic integration. We validate our method to compute solvation free energies on several benchmark systems: a Lennard-Jones particle in a Lennard-Jones fluid, as well as the insertion of both water and methane solutes in a water solvent at atomistic resolution.

☆ Mean-Field Simulation-Based Inference for Cosmological Initial Conditions NeurIPS 2024

Oleg Savchenko, Florian List, Guillermo Franco Abellán, Noemi Anau Montel, Christoph Weniger

Reconstructing cosmological initial conditions (ICs) from late-time observations is a difficult task, which relies on the use of computationally expensive simulators alongside sophisticated statistical methods to navigate multi-million dimensional parameter spaces. We present a simple method for Bayesian field reconstruction based on modeling the posterior distribution of the initial matter density field to be diagonal Gaussian in Fourier space, with its covariance and the mean estimator being the trainable parts of the algorithm. Training and sampling are extremely fast (training: $\sim 1 \, \mathrm{h}$ on a GPU, sampling: $\lesssim 3 \, \mathrm{s}$ for 1000 samples at resolution $128^3$), and our method supports industry-standard (non-differentiable) $N$-body simulators. We verify the fidelity of the obtained IC samples in terms of summary statistics.

comment: Accepted for the NeurIPS 2024 workshop Machine Learning and the Physical Sciences; 5 + 4 pages, 3 figures

☆ Deep Learning and Data Augmentation for Detecting Self-Admitted Technical Debt

Edi Sutoyo, Paris Avgeriou, Andrea Capiluppi

Self-Admitted Technical Debt (SATD) refers to circumstances where developers use textual artifacts to explain why the existing implementation is not optimal. Past research in detecting SATD has focused on either identifying SATD (classifying SATD items as SATD or not) or categorizing SATD (labeling instances as SATD that pertain to requirement, design, code, test debt, etc.). However, the performance of these approaches remains suboptimal, particularly for specific types of SATD, such as test and requirement debt, primarily due to extremely imbalanced datasets. To address these challenges, we build on earlier research by utilizing BiLSTM architecture for the binary identification of SATD and BERT architecture for categorizing different types of SATD. Despite their effectiveness, both architectures struggle with imbalanced data. Therefore, we employ a large language model data augmentation strategy to mitigate this issue. Furthermore, we introduce a two-step approach to identify and categorize SATD across various datasets derived from different artifacts. Our contributions include providing a balanced dataset for future SATD researchers and demonstrating that our approach significantly improves SATD identification and categorization performance compared to baseline methods.

comment: Accepted to be published at the 2024 31st Asia-Pacific Software Engineering Conference (APSEC)

☆ On the VC dimension of deep group convolutional neural networks

Anna Sepliarskaia, Sophie Langer, Johannes Schmidt-Hieber

We study the generalization capabilities of Group Convolutional Neural Networks (GCNNs) with ReLU activation function by deriving upper and lower bounds for their Vapnik-Chervonenkis (VC) dimension. Specifically, we analyze how factors such as the number of layers, weights, and input dimension affect the VC dimension. We further compare the derived bounds to those known for other types of neural networks. Our findings extend previous results on the VC dimension of continuous GCNNs with two layers, thereby providing new insights into the generalization properties of GCNNs, particularly regarding the dependence on the input resolution of the data.

☆ Arithmetic Transformers Can Length-Generalize in Both Operand Length and Count

Hanseul Cho, Jaeyoung Cha, Srinadh Bhojanapalli, Chulhee Yun

Transformers often struggle with length generalization, meaning they fail to generalize to sequences longer than those encountered during training. While arithmetic tasks are commonly used to study length generalization, certain tasks are considered notoriously difficult, e.g., multi-operand addition (requiring generalization over both the number of operands and their lengths) and multiplication (requiring generalization over both operand lengths). In this work, we achieve approximately 2-3x length generalization on both tasks, which is the first such achievement in arithmetic Transformers. We design task-specific scratchpads enabling the model to focus on a fixed number of tokens per each next-token prediction step, and apply multi-level versions of Position Coupling (Cho et al., 2024; McLeish et al., 2024) to let Transformers know the right position to attend to. On the theory side, we prove that a 1-layer Transformer using our method can solve multi-operand addition, up to operand length and operand count that are exponential in embedding dimension.

comment: 38 pages, 16 figures

☆ Reducing Hallucinations in Vision-Language Models via Latent Space Steering

Sheng Liu, Haotian Ye, James Zou

Hallucination poses a challenge to the deployment of large vision-language models (LVLMs) in applications. Unlike in large language models (LLMs), hallucination in LVLMs often arises from misalignments between visual inputs and textual outputs. This paper investigates the underlying mechanisms of hallucination, focusing on the unique structure of LVLMs that distinguishes them from large language models (LLMs). We identify that hallucinations often arise from the sensitivity of text decoders to vision inputs, a natural phenomenon when image encoders and text decoders are pre-trained separately. Inspired by this, we introduce Visual and Textual Intervention (VTI), a novel technique designed to reduce hallucinations by steering latent space representations during inference to enhance the stability of vision features. As a task-agnostic test-time intervention, VTI can be easily applied to any problem without additional cost. Extensive experiments demonstrate that it can effectively reduce hallucinations and outperform baseline methods across multiple metrics, highlighting the critical role of vision feature stability in LVLMs.

comment: 21 pages

☆ High-Fidelity Transfer of Functional Priors for Wide Bayesian Neural Networks by Learning Activations

Marcin Sendera, Amin Sorkhei, Tomasz Kuśmierczyk

Function-space priors in Bayesian Neural Networks provide a more intuitive approach to embedding beliefs directly into the model's output, thereby enhancing regularization, uncertainty quantification, and risk-aware decision-making. However, imposing function-space priors on BNNs is challenging. We address this task through optimization techniques that explore how trainable activations can accommodate complex priors and match intricate target function distributions. We discuss critical learning challenges, including identifiability, loss construction, and symmetries that arise in this context. Furthermore, we enable evidence maximization to facilitate model selection by conditioning the functional priors on additional hyperparameters. Our empirical findings demonstrate that even BNNs with a single wide hidden layer, when equipped with these adaptive trainable activations and conditioning strategies, can effectively achieve high-fidelity function-space priors, providing a robust and flexible framework for enhancing Bayesian neural network performance.

☆ Mislabeled examples detection viewed as probing machine learning models: concepts, survey and extensive benchmark

Thomas George, Pierre Nodet, Alexis Bondu, Vincent Lemaire

Mislabeled examples are ubiquitous in real-world machine learning datasets, advocating the development of techniques for automatic detection. We show that most mislabeled detection methods can be viewed as probing trained machine learning models using a few core principles. We formalize a modular framework that encompasses these methods, parameterized by only 4 building blocks, as well as a Python library that demonstrates that these principles can actually be implemented. The focus is on classifier-agnostic concepts, with an emphasis on adapting methods developed for deep learning models to non-deep classifiers for tabular data. We benchmark existing methods on (artificial) Completely At Random (NCAR) as well as (realistic) Not At Random (NNAR) labeling noise from a variety of tasks with imperfect labeling rules. This benchmark provides new insights as well as limitations of existing methods in this setup.

☆ SeisLM: a Foundation Model for Seismic Waveforms

Tianlin Liu, Jannes Münchmeyer, Laura Laurenti, Chris Marone, Maarten V. de Hoop, Ivan Dokmanić

We introduce the Seismic Language Model (SeisLM), a foundational model designed to analyze seismic waveforms -- signals generated by Earth's vibrations such as the ones originating from earthquakes. SeisLM is pretrained on a large collection of open-source seismic datasets using a self-supervised contrastive loss, akin to BERT in language modeling. This approach allows the model to learn general seismic waveform patterns from unlabeled data without being tied to specific downstream tasks. When fine-tuned, SeisLM excels in seismological tasks like event detection, phase-picking, onset time regression, and foreshock-aftershock classification. The code has been made publicly available on https://github.com/liutianlin0121/seisLM.

☆ Solving Sparse \& High-Dimensional-Output Regression via Compression

Renyuan Li, Zhehui Chen, Guanyi Wang

Multi-Output Regression (MOR) has been widely used in scientific data analysis for decision-making. Unlike traditional regression models, MOR aims to simultaneously predict multiple real-valued outputs given an input. However, the increasing dimensionality of the outputs poses significant challenges regarding interpretability and computational scalability for modern MOR applications. As a first step to address these challenges, this paper proposes a Sparse \& High-dimensional-Output REgression (SHORE) model by incorporating additional sparsity requirements to resolve the output interpretability, and then designs a computationally efficient two-stage optimization framework capable of solving SHORE with provable accuracy via compression on outputs. Theoretically, we show that the proposed framework is computationally scalable while maintaining the same order of training loss and prediction loss before-and-after compression under arbitrary or relatively weak sample set conditions. Empirically, numerical results further validate the theoretical findings, showcasing the efficiency and accuracy of the proposed framework.

comment: Admitted in Neurips 2024

☆ Learning-to-Defer for Extractive Question Answering

Montreuil Yannis, Carlier Axel, Ng Lai Xing, Ooi Wei Tsang

Pre-trained language models have profoundly impacted the field of extractive question-answering, leveraging large-scale textual corpora to enhance contextual language understanding. Despite their success, these models struggle in complex scenarios that demand nuanced interpretation or inferential reasoning beyond immediate textual cues. Furthermore, their size poses deployment challenges on resource-constrained devices. Addressing these limitations, we introduce an adapted two-stage Learning-to-Defer mechanism that enhances decision-making by enabling selective deference to human experts or larger models without retraining language models in the context of question-answering. This approach not only maintains computational efficiency but also significantly improves model reliability and accuracy in ambiguous contexts. We establish the theoretical soundness of our methodology by proving Bayes and $(\mathcal{H}, \mathcal{R})$--consistency of our surrogate loss function, guaranteeing the optimality of the final solution. Empirical evaluations on the SQuADv2 dataset illustrate performance gains from integrating human expertise and leveraging larger models. Our results further demonstrate that deferring a minimal number of queries allows the smaller model to achieve performance comparable to their larger counterparts while preserving computing efficiency, thus broadening the applicability of pre-trained language models in diverse operational environments.

comment: 25 pages, 17 main paper

☆ DeepVigor+: Scalable and Accurate Semi-Analytical Fault Resilience Analysis for Deep Neural Network

Mohammad Hasan Ahmadilivani, Jaan Raik, Masoud Daneshtalab, Maksim Jenihhin

Growing exploitation of Machine Learning (ML) in safety-critical applications necessitates rigorous safety analysis. Hardware reliability assessment is a major concern with respect to measuring the level of safety. Quantifying the reliability of emerging ML models, including Deep Neural Networks (DNNs), is highly complex due to their enormous size in terms of the number of parameters and computations. Conventionally, Fault Injection (FI) is applied to perform a reliability measurement. However, performing FI on modern-day DNNs is prohibitively time-consuming if an acceptable confidence level is to be achieved. In order to speed up FI for large DNNs, statistical FI has been proposed. However, the run-time for the large DNN models is still considerably long. In this work, we introduce DeepVigor+, a scalable, fast and accurate semi-analytical method as an efficient alternative for reliability measurement in DNNs. DeepVigor+ implements a fault propagation analysis model and attempts to acquire Vulnerability Factors (VFs) as reliability metrics in an optimal way. The results indicate that DeepVigor+ obtains VFs for DNN models with an error less than 1\% and 14.9 up to 26.9 times fewer simulations than the best-known state-of-the-art statistical FI enabling an accurate reliability analysis for emerging DNNs within a few minutes.

comment: 14 pages, 9 figures, 8 tables, 16 equations. The source code is accessible via: https://github.com/mhahmadilivany/DeepVigor

☆ Two-stage Learning-to-Defer for Multi-Task Learning

Montreuil Yannis, Yeo Shu Heng, Carlier Axel, Ng Lai Xing, Ooi Wei Tsang

The Learning-to-Defer approach has been explored for classification and, more recently, regression tasks separately. Many contemporary learning tasks, however, involves both classification and regression components. In this paper, we introduce a Learning-to-Defer approach for multi-task learning that encompasses both classification and regression tasks. Our two-stage approach utilizes a rejector that defers decisions to the most accurate agent among a pre-trained joint classifier-regressor models and one or more external experts. We show that our surrogate loss is $(\mathcal{H}, \mathcal{F}, \mathcal{R})$ and Bayes--consistent, ensuring an effective approximation of the optimal solution. Additionally, we derive learning bounds that demonstrate the benefits of employing multiple confident experts along a rich model in a two-stage learning framework. Empirical experiments conducted on electronic health record analysis tasks underscore the performance enhancements achieved through our method.

comment: 32 pages, 17 main paper

★ Object-Centric Temporal Consistency via Conditional Autoregressive Inductive Biases

Cristian Meo, Akihiro Nakano, Mircea Lică, Aniket Didolkar, Masahiro Suzuki, Anirudh Goyal, Mengmi Zhang, Justin Dauwels, Yutaka Matsuo, Yoshua Bengio

Unsupervised object-centric learning from videos is a promising approach towards learning compositional representations that can be applied to various downstream tasks, such as prediction and reasoning. Recently, it was shown that pretrained Vision Transformers (ViTs) can be useful to learn object-centric representations on real-world video datasets. However, while these approaches succeed at extracting objects from the scenes, the slot-based representations fail to maintain temporal consistency across consecutive frames in a video, i.e. the mapping of objects to slots changes across the video. To address this, we introduce Conditional Autoregressive Slot Attention (CA-SA), a framework that enhances the temporal consistency of extracted object-centric representations in video-centric vision tasks. Leveraging an autoregressive prior network to condition representations on previous timesteps and a novel consistency loss function, CA-SA predicts future slot representations and imposes consistency across frames. We present qualitative and quantitative results showing that our proposed method outperforms the considered baselines on downstream tasks, such as video prediction and visual question-answering tasks.

☆ S-CFE: Simple Counterfactual Explanations

Shpresim Sadiku, Moritz Wagner, Sai Ganesh Nagarajan, Sebastian Pokutta

We study the problem of finding optimal sparse, manifold-aligned counterfactual explanations for classifiers. Canonically, this can be formulated as an optimization problem with multiple non-convex components, including classifier loss functions and manifold alignment (or \emph{plausibility}) metrics. The added complexity of enforcing \emph{sparsity}, or shorter explanations, complicates the problem further. Existing methods often focus on specific models and plausibility measures, relying on convex $\ell_1$ regularizers to enforce sparsity. In this paper, we tackle the canonical formulation using the accelerated proximal gradient (APG) method, a simple yet efficient first-order procedure capable of handling smooth non-convex objectives and non-smooth $\ell_p$ (where $0 \leq p < 1$) regularizers. This enables our approach to seamlessly incorporate various classifiers and plausibility measures while producing sparser solutions. Our algorithm only requires differentiable data-manifold regularizers and supports box constraints for bounded feature ranges, ensuring the generated counterfactuals remain \emph{actionable}. Finally, experiments on real-world datasets demonstrate that our approach effectively produces sparse, manifold-aligned counterfactual explanations while maintaining proximity to the factual data and computational efficiency.

☆ Learning signals defined on graphs with optimal transport and Gaussian process regression

Raphaël Carpintero Perez, Sébastien da Veiga, Josselin Garnier, Brian Staber

In computational physics, machine learning has now emerged as a powerful complementary tool to explore efficiently candidate designs in engineering studies. Outputs in such supervised problems are signals defined on meshes, and a natural question is the extension of general scalar output regression models to such complex outputs. Changes between input geometries in terms of both size and adjacency structure in particular make this transition non-trivial. In this work, we propose an innovative strategy for Gaussian process regression where inputs are large and sparse graphs with continuous node attributes and outputs are signals defined on the nodes of the associated inputs. The methodology relies on the combination of regularized optimal transport, dimension reduction techniques, and the use of Gaussian processes indexed by graphs. In addition to enabling signal prediction, the main point of our proposal is to come with confidence intervals on node values, which is crucial for uncertainty quantification and active learning. Numerical experiments highlight the efficiency of the method to solve real problems in fluid dynamics and solid mechanics.

☆ Traffic Matrix Estimation based on Denoising Diffusion Probabilistic Model

Xinyu Yuan, Yan Qiao, Pei Zhao, Rongyao Hu, Benchu Zhang

The traffic matrix estimation (TME) problem has been widely researched for decades of years. Recent progresses in deep generative models offer new opportunities to tackle TME problems in a more advanced way. In this paper, we leverage the powerful ability of denoising diffusion probabilistic models (DDPMs) on distribution learning, and for the first time adopt DDPM to address the TME problem. To ensure a good performance of DDPM on learning the distributions of TMs, we design a preprocessing module to reduce the dimensions of TMs while keeping the data variety of each OD flow. To improve the estimation accuracy, we parameterize the noise factors in DDPM and transform the TME problem into a gradient-descent optimization problem. Finally, we compared our method with the state-of-the-art TME methods using two real-world TM datasets, the experimental results strongly demonstrate the superiority of our method on both TM synthesis and TM estimation.

☆ Offline reinforcement learning for job-shop scheduling problems

Imanol Echeverria, Maialen Murua, Roberto Santana

Recent advances in deep learning have shown significant potential for solving combinatorial optimization problems in real-time. Unlike traditional methods, deep learning can generate high-quality solutions efficiently, which is crucial for applications like routing and scheduling. However, existing approaches like deep reinforcement learning (RL) and behavioral cloning have notable limitations, with deep RL suffering from slow learning and behavioral cloning relying solely on expert actions, which can lead to generalization issues and neglect of the optimization objective. This paper introduces a novel offline RL method designed for combinatorial optimization problems with complex constraints, where the state is represented as a heterogeneous graph and the action space is variable. Our approach encodes actions in edge attributes and balances expected rewards with the imitation of expert solutions. We demonstrate the effectiveness of this method on job-shop scheduling and flexible job-shop scheduling benchmarks, achieving superior performance compared to state-of-the-art techniques.

☆ Estimating Individual Dose-Response Curves under Unobserved Confounders from Observational Data

Shutong Chen, Yang Li

Estimating an individual's potential response to continuously varied treatments is crucial for addressing causal questions across diverse domains, from healthcare to social sciences. However, existing methods are limited either to estimating causal effects of binary treatments, or scenarios where all confounding variables are measurable. In this work, we present ContiVAE, a novel framework for estimating causal effects of continuous treatments, measured by individual dose-response curves, considering the presence of unobserved confounders using observational data. Leveraging a variational auto-encoder with a Tilted Gaussian prior distribution, ContiVAE models the hidden confounders as latent variables, and is able to predict the potential outcome of any treatment level for each individual while effectively capture the heterogeneity among individuals. Experiments on semi-synthetic datasets show that ContiVAE outperforms existing methods by up to 62%, demonstrating its robustness and flexibility. Application on a real-world dataset illustrates its practical utility.

☆ Residual vector quantization for KV cache compression in large language model

Ankur Kumar

KV cache compression methods have mainly relied on scalar quantization techniques to reduce the memory requirements during decoding. In this work, we apply residual vector quantization, which has been widely used for high fidelity audio compression, to compress KV cache in large language models (LLM). We adapt the standard recipe with minimal changes to compress the output of any key or value projection matrix in a pretrained LLM: we scale the vector by its standard deviation, divide channels into groups and then quantize each group with the same residual vector quantizer. We learn the codebook using exponential moving average and there are no other learnable parameters including the input and output projections normally used in a vector quantization set up. We find that a residual depth of 8 recovers most of the performance of the unquantized model. We also find that grouping non-contiguous channels together works better than grouping contiguous channels for compressing key matrix and the method further benefits from a light weight finetuning of LLM together with the quantization. Overall, the proposed technique is competitive with existing quantization methods while being much simpler and results in 5.5x compression compared to half precision.

☆ Solving Continual Offline RL through Selective Weights Activation on Aligned Spaces

Jifeng Hu, Sili Huang, Li Shen, Zhejian Yang, Shengchao Hu, Shisong Tang, Hechang Chen, Yi Chang, Dacheng Tao, Lichao Sun

Continual offline reinforcement learning (CORL) has shown impressive ability in diffusion-based lifelong learning systems by modeling the joint distributions of trajectories. However, most research only focuses on limited continual task settings where the tasks have the same observation and action space, which deviates from the realistic demands of training agents in various environments. In view of this, we propose Vector-Quantized Continual Diffuser, named VQ-CD, to break the barrier of different spaces between various tasks. Specifically, our method contains two complementary sections, where the quantization spaces alignment provides a unified basis for the selective weights activation. In the quantized spaces alignment, we leverage vector quantization to align the different state and action spaces of various tasks, facilitating continual training in the same space. Then, we propose to leverage a unified diffusion model attached by the inverse dynamic model to master all tasks by selectively activating different weights according to the task-related sparse masks. Finally, we conduct extensive experiments on 15 continual learning (CL) tasks, including conventional CL task settings (identical state and action spaces) and general CL task settings (various state and action spaces). Compared with 16 baselines, our method reaches the SOTA performance.

☆ Enhancing SNN-based Spatio-Temporal Learning: A Benchmark Dataset and Cross-Modality Attention Model

Shibo Zhou, Bo Yang, Mengwen Yuan, Runhao Jiang, Rui Yan, Gang Pan, Huajin Tang

Spiking Neural Networks (SNNs), renowned for their low power consumption, brain-inspired architecture, and spatio-temporal representation capabilities, have garnered considerable attention in recent years. Similar to Artificial Neural Networks (ANNs), high-quality benchmark datasets are of great importance to the advances of SNNs. However, our analysis indicates that many prevalent neuromorphic datasets lack strong temporal correlation, preventing SNNs from fully exploiting their spatio-temporal representation capabilities. Meanwhile, the integration of event and frame modalities offers more comprehensive visual spatio-temporal information. Yet, the SNN-based cross-modality fusion remains underexplored. In this work, we present a neuromorphic dataset called DVS-SLR that can better exploit the inherent spatio-temporal properties of SNNs. Compared to existing datasets, it offers advantages in terms of higher temporal correlation, larger scale, and more varied scenarios. In addition, our neuromorphic dataset contains corresponding frame data, which can be used for developing SNN-based fusion methods. By virtue of the dual-modal feature of the dataset, we propose a Cross-Modality Attention (CMA) based fusion method. The CMA model efficiently utilizes the unique advantages of each modality, allowing for SNNs to learn both temporal and spatial attention scores from the spatio-temporal features of event and frame modalities, subsequently allocating these scores across modalities to enhance their synergy. Experimental results demonstrate that our method not only improves recognition accuracy but also ensures robustness across diverse scenarios.

☆ MIK: Modified Isolation Kernel for Biological Sequence Visualization, Classification, and Clustering

Sarwan Ali, Prakash Chourasia, Haris Mansoor, Bipin koirala, Murray Patterson

The t-Distributed Stochastic Neighbor Embedding (t-SNE) has emerged as a popular dimensionality reduction technique for visualizing high-dimensional data. It computes pairwise similarities between data points by default using an RBF kernel and random initialization (in low-dimensional space), which successfully captures the overall structure but may struggle to preserve the local structure efficiently. This research proposes a novel approach called the Modified Isolation Kernel (MIK) as an alternative to the Gaussian kernel, which is built upon the concept of the Isolation Kernel. MIK uses adaptive density estimation to capture local structures more accurately and integrates robustness measures. It also assigns higher similarity values to nearby points and lower values to distant points. Comparative research using the normal Gaussian kernel, the isolation kernel, and several initialization techniques, including random, PCA, and random walk initializations, are used to assess the proposed approach (MIK). Additionally, we compare the computational efficiency of all $3$ kernels with $3$ different initialization methods. Our experimental results demonstrate several advantages of the proposed kernel (MIK) and initialization method selection. It exhibits improved preservation of the local and global structure and enables better visualization of clusters and subclusters in the embedded space. These findings contribute to advancing dimensionality reduction techniques and provide researchers and practitioners with an effective tool for data exploration, visualization, and analysis in various domains.

☆ Federated Learning with MMD-based Early Stopping for Adaptive GNSS Interference Classification

Nishant S. Gaikwad, Lucas Heublein, Nisha L. Raichur, Tobias Feigl, Christopher Mutschler, Felix Ott

Federated learning (FL) enables multiple devices to collaboratively train a global model while maintaining data on local servers. Each device trains the model on its local server and shares only the model updates (i.e., gradient weights) during the aggregation step. A significant challenge in FL is managing the feature distribution of novel, unbalanced data across devices. In this paper, we propose an FL approach using few-shot learning and aggregation of the model weights on a global server. We introduce a dynamic early stopping method to balance out-of-distribution classes based on representation learning, specifically utilizing the maximum mean discrepancy of feature embeddings between local and global models. An exemplary application of FL is orchestrating machine learning models along highways for interference classification based on snapshots from global navigation satellite system (GNSS) receivers. Extensive experiments on four GNSS datasets from two real-world highways and controlled environments demonstrate that our FL method surpasses state-of-the-art techniques in adapting to both novel interference classes and multipath scenarios.

☆ RAC: Efficient LLM Factuality Correction with Retrieval Augmentation

Changmao Li, Jeffrey Flanigan

Large Language Models (LLMs) exhibit impressive results across a wide range of natural language processing (NLP) tasks, yet they can often produce factually incorrect outputs. This paper introduces a simple but effective low-latency post-correction method, \textbf{Retrieval Augmented Correction (RAC)}, aimed at enhancing the factual performance of LLMs without requiring additional fine-tuning. Our method is general and can be used with any instruction-tuned LLM, and has greatly reduced latency compared to prior approaches. RAC decomposes the LLM's output into atomic facts and applies a fine-grained verification and correction process with retrieved content to verify and correct the LLM-generated output. Our extensive experiments show that RAC yields up to 30\% improvements over state-of-the-art baselines across two popular factuality evaluation datasets, validating its efficacy and robustness in both with and without the integration of Retrieval-Augmented Generation (RAG) across different LLMs.\footnote{Our code is at \url{https://github.com/jlab-nlp/Retrieval-Augmented-Correction}}

☆ Long Term Memory: The Foundation of AI Self-Evolution

Xun Jiang, Feng Li, Han Zhao, Jiaying Wang, Jun Shao, Shihao Xu, Shu Zhang, Weiling Chen, Xavier Tang, Yize Chen, Mengyue Wu, Weizhi Ma, Mengdi Wang, Tianqiao Chen

Large language models (LLMs) like GPTs, trained on vast datasets, have demonstrated impressive capabilities in language understanding, reasoning, and planning, achieving human-level performance in various tasks. Most studies focus on enhancing these models by training on ever-larger datasets to build more powerful foundation models. While training stronger models is important, enabling models to evolve during inference is equally crucial, a process we refer to as AI self-evolution. Unlike large-scale training, self-evolution may rely on limited data or interactions. Inspired by the columnar organization of the human cerebral cortex, we hypothesize that AI models could develop cognitive abilities and build internal representations through iterative interactions with their environment. To achieve this, models need long-term memory (LTM) to store and manage processed interaction data. LTM supports self-evolution by representing diverse experiences across environments and agents. In this report, we explore AI self-evolution and its potential to enhance models during inference. We examine LTM's role in lifelong learning, allowing models to evolve based on accumulated interactions. We outline the structure of LTM and the systems needed for effective data retention and representation. We also classify approaches for building personalized models with LTM data and show how these models achieve self-evolution through interaction. Using LTM, our multi-agent framework OMNE achieved first place on the GAIA benchmark, demonstrating LTM's potential for AI self-evolution. Finally, we present a roadmap for future research, emphasizing the importance of LTM for advancing AI technology and its practical applications.

comment: 56 pages, 13 figures

☆ Scalable Data Ablation Approximations for Language Models through Modular Training and Merging EMNLP 2024

Clara Na, Ian Magnusson, Ananya Harsh Jha, Tom Sherborne, Emma Strubell, Jesse Dodge, Pradeep Dasigi

Training data compositions for Large Language Models (LLMs) can significantly affect their downstream performance. However, a thorough data ablation study exploring large sets of candidate data mixtures is typically prohibitively expensive since the full effect is seen only after training the models; this can lead practitioners to settle for sub-optimal data mixtures. We propose an efficient method for approximating data ablations which trains individual models on subsets of a training corpus and reuses them across evaluations of combinations of subsets. In continued pre-training experiments, we find that, given an arbitrary evaluation set, the perplexity score of a single model trained on a candidate set of data is strongly correlated with perplexity scores of parameter averages of models trained on distinct partitions of that data. From this finding, we posit that researchers and practitioners can conduct inexpensive simulations of data ablations by maintaining a pool of models that were each trained on partitions of a large training corpus, and assessing candidate data mixtures by evaluating parameter averages of combinations of these models. This approach allows for substantial improvements in amortized training efficiency -- scaling only linearly with respect to new data -- by enabling reuse of previous training computation, opening new avenues for improving model performance through rigorous, incremental data assessment and mixing.

comment: EMNLP 2024. 17 pages

☆ Calibration of ordinal regression networks

Daehwan Kim, Haejun Chung, Ikbeom Jang

Recent studies have shown that deep neural networks are not well-calibrated and produce over-confident predictions. The miscalibration issue primarily stems from the minimization of cross-entropy, which aims to align predicted softmax probabilities with one-hot labels. In ordinal regression tasks, this problem is compounded by an additional challenge: the expectation that softmax probabilities should exhibit unimodal distribution is not met with cross-entropy. Rather, the ordinal regression literature has focused on unimodality and overlooked calibration. To address these issues, we propose a novel loss function that introduces order-aware calibration, ensuring that prediction confidence adheres to ordinal relationships between classes. It incorporates soft ordinal encoding and label-smoothing-based regularization to enforce both calibration and unimodality. Extensive experiments across three popular ordinal regression benchmarks demonstrate that our approach achieves state-of-the-art calibration without compromising accuracy.

☆ Accounting for Missing Covariates in Heterogeneous Treatment Estimation

Khurram Yamin, Vibhhu Sharma, Ed Kennedy, Bryan Wilder

Many applications of causal inference require using treatment effects estimated on a study population to make decisions in a separate target population. We consider the challenging setting where there are covariates that are observed in the target population that were not seen in the original study. Our goal is to estimate the tightest possible bounds on heterogeneous treatment effects conditioned on such newly observed covariates. We introduce a novel partial identification strategy based on ideas from ecological inference; the main idea is that estimates of conditional treatment effects for the full covariate set must marginalize correctly when restricted to only the covariates observed in both populations. Furthermore, we introduce a bias-corrected estimator for these bounds and prove that it enjoys fast convergence rates and statistical guarantees (e.g., asymptotic normality). Experimental results on both real and synthetic data demonstrate that our framework can produce bounds that are much tighter than would otherwise be possible.

☆ Understanding and Alleviating Memory Consumption in RLHF for LLMs

Jin Zhou, Hanmei Yang, Steven, Tang, Mingcan Xiang, Hui Guan, Tongping Liu

Fine-tuning with Reinforcement Learning with Human Feedback (RLHF) is essential for aligning large language models (LLMs). However, RLHF often encounters significant memory challenges. This study is the first to examine memory usage in the RLHF context, exploring various memory management strategies and unveiling the reasons behind excessive memory consumption. Additionally, we introduce a simple yet effective approach that substantially reduces the memory required for RLHF fine-tuning.

☆ Linking Model Intervention to Causal Interpretation in Model Explanation

Debo Cheng, Ziqi Xu, Jiuyong Li, Lin Liu, Kui Yu, Thuc Duy Le, Jixue Liu

Intervention intuition is often used in model explanation where the intervention effect of a feature on the outcome is quantified by the difference of a model prediction when the feature value is changed from the current value to the baseline value. Such a model intervention effect of a feature is inherently association. In this paper, we will study the conditions when an intuitive model intervention effect has a causal interpretation, i.e., when it indicates whether a feature is a direct cause of the outcome. This work links the model intervention effect to the causal interpretation of a model. Such an interpretation capability is important since it indicates whether a machine learning model is trustworthy to domain experts. The conditions also reveal the limitations of using a model intervention effect for causal interpretation in an environment with unobserved features. Experiments on semi-synthetic datasets have been conducted to validate theorems and show the potential for using the model intervention effect for model interpretation.

☆ Deep Graph Attention Networks

Jun Kato, Airi Mita, Keita Gobara, Akihiro Inokuchi

Graphs are useful for representing various realworld objects. However, graph neural networks (GNNs) tend to suffer from over-smoothing, where the representations of nodes of different classes become similar as the number of layers increases, leading to performance degradation. A method that does not require protracted tuning of the number of layers is needed to effectively construct a graph attention network (GAT), a type of GNN. Therefore, we introduce a method called "DeepGAT" for predicting the class to which nodes belong in a deep GAT. It avoids over-smoothing in a GAT by ensuring that nodes in different classes are not similar at each layer. Using DeepGAT to predict class labels, a 15-layer network is constructed without the need to tune the number of layers. DeepGAT prevented over-smoothing and achieved a 15-layer GAT with similar performance to a 2-layer GAT, as indicated by the similar attention coefficients. DeepGAT enables the training of a large network to acquire similar attention coefficients to a network with few layers. It avoids the over-smoothing problem and obviates the need to tune the number of layers, thus saving time and enhancing GNN performance.

comment: 8 pages, 6 figures

☆ Large Deviations and Improved Mean-squared Error Rates of Nonlinear SGD: Heavy-tailed Noise and Power of Symmetry

Aleksandar Armacki, Shuhua Yu, Dragana Bajovic, Dusan Jakovetic, Soummya Kar

We study large deviations and mean-squared error (MSE) guarantees of a general framework of nonlinear stochastic gradient methods in the online setting, in the presence of heavy-tailed noise. Unlike existing works that rely on the closed form of a nonlinearity (typically clipping), our framework treats the nonlinearity in a black-box manner, allowing us to provide unified guarantees for a broad class of bounded nonlinearities, including many popular ones, like sign, quantization, normalization, as well as component-wise and joint clipping. We provide several strong results for a broad range of step-sizes in the presence of heavy-tailed noise with symmetric probability density function, positive in a neighbourhood of zero and potentially unbounded moments. In particular, for non-convex costs we provide a large deviation upper bound for the minimum norm-squared of gradients, showing an asymptotic tail decay on an exponential scale, at a rate $\sqrt{t} / \log(t)$. We establish the accompanying rate function, showing an explicit dependence on the choice of step-size, nonlinearity, noise and problem parameters. Next, for non-convex costs and the minimum norm-squared of gradients, we derive the optimal MSE rate $\widetilde{\mathcal{O}}(t^{-1/2})$. Moreover, for strongly convex costs and the last iterate, we provide an MSE rate that can be made arbitrarily close to the optimal rate $\mathcal{O}(t^{-1})$, improving on the state-of-the-art results in the presence of heavy-tailed noise. Finally, we establish almost sure convergence of the minimum norm-squared of gradients, providing an explicit rate, which can be made arbitrarily close to $o(t^{-1/4})$.

comment: 30 pages. arXiv admin note: text overlap with arXiv:2410.13954

☆ Towards Kriging-informed Conditional Diffusion for Regional Sea-Level Data Downscaling

Subhankar Ghosh, Arun Sharma, Jayant Gupta, Aneesh Subramanian, Shashi Shekhar

Given coarser-resolution projections from global climate models or satellite data, the downscaling problem aims to estimate finer-resolution regional climate data, capturing fine-scale spatial patterns and variability. Downscaling is any method to derive high-resolution data from low-resolution variables, often to provide more detailed and local predictions and analyses. This problem is societally crucial for effective adaptation, mitigation, and resilience against significant risks from climate change. The challenge arises from spatial heterogeneity and the need to recover finer-scale features while ensuring model generalization. Most downscaling methods \cite{Li2020} fail to capture the spatial dependencies at finer scales and underperform on real-world climate datasets, such as sea-level rise. We propose a novel Kriging-informed Conditional Diffusion Probabilistic Model (Ki-CDPM) to capture spatial variability while preserving fine-scale features. Experimental results on climate data show that our proposed method is more accurate than state-of-the-art downscaling techniques.

☆ Improving Parallel Program Performance Through DSL-Driven Code Generation with LLM Optimizers

Anjiang Wei, Allen Nie, Thiago S. F. X. Teixeira, Rohan Yadav, Wonchan Lee, Ke Wang, Alex Aiken

Mapping computations to processors and assigning data to memory are critical for maximizing performance in parallel programming. These mapping decisions are managed through the development of specialized low-level system code, called mappers, crafted by performance engineers. Each mapper is tailored to a specific application and optimized for the underlying machine architecture, a process that requires days of refinement and tuning from an expert. Despite advances in system research, automating mapper generation remains a challenge due to the complexity of making millions of decisions to find the optimal solution and generate the solution as code. We introduce an approach that leverages recent advances in LLM-based optimizers for mapper design. In under ten minutes, our method automatically discovers mappers that surpass human expert designs in scientific applications by up to 1.34X speedup. For parallel matrix multiplication algorithms, our mapper achieves up to 1.31X of the expert-designed solution. To achieve this, we simplify the complexity of low-level code generation by introducing a domain-specific language (DSL) that abstracts the low-level system programming details and defines a structured search space for LLMs to explore. To maximize the application performance, we use an LLM optimizer to improve an agentic system that generates the mapper code. As a result, this approach significantly reduces the workload for performance engineers while achieving substantial performance gains across diverse applications. Finally, our results demonstrate the effectiveness of LLM-based optimization in system design and suggest its potential for addressing other complex system challenges.

comment: 26 pages, 8 figures

☆ Test-time Adaptation for Cross-modal Retrieval with Query Shift

Haobin Li, Peng Hu, Qianjun Zhang, Xi Peng, Xiting Liu, Mouxing Yang

The success of most existing cross-modal retrieval methods heavily relies on the assumption that the given queries follow the same distribution of the source domain. However, such an assumption is easily violated in real-world scenarios due to the complexity and diversity of queries, thus leading to the query shift problem. Specifically, query shift refers to the online query stream originating from the domain that follows a different distribution with the source one. In this paper, we observe that query shift would not only diminish the uniformity (namely, within-modality scatter) of the query modality but also amplify the gap between query and gallery modalities. Based on the observations, we propose a novel method dubbed Test-time adaptation for Cross-modal Retrieval (TCR). In brief, TCR employs a novel module to refine the query predictions (namely, retrieval results of the query) and a joint objective to prevent query shift from disturbing the common space, thus achieving online adaptation for the cross-modal retrieval models with query shift. Expensive experiments demonstrate the effectiveness of the proposed TCR against query shift. The code will be released upon acceptance.

comment: 22 pages, 8 figures

☆ Erasing Undesirable Concepts in Diffusion Models with Adversarial Preservation

Anh Bui, Long Vuong, Khanh Doan, Trung Le, Paul Montague, Tamas Abraham, Dinh Phung

Diffusion models excel at generating visually striking content from text but can inadvertently produce undesirable or harmful content when trained on unfiltered internet data. A practical solution is to selectively removing target concepts from the model, but this may impact the remaining concepts. Prior approaches have tried to balance this by introducing a loss term to preserve neutral content or a regularization term to minimize changes in the model parameters, yet resolving this trade-off remains challenging. In this work, we propose to identify and preserving concepts most affected by parameter changes, termed as \textit{adversarial concepts}. This approach ensures stable erasure with minimal impact on the other concepts. We demonstrate the effectiveness of our method using the Stable Diffusion model, showing that it outperforms state-of-the-art erasure methods in eliminating unwanted content while maintaining the integrity of other unrelated elements. Our code is available at \url{https://github.com/tuananhbui89/Erasing-Adversarial-Preservation}.

☆ Long-time Integration of Nonlinear Wave Equations with Neural Operators

Guanhang Lei, Zhen Lei, Lei Shi

Neural operators have shown promise in solving many types of Partial Differential Equations (PDEs). They are significantly faster compared to traditional numerical solvers once they have been trained with a certain amount of observed data. However, their numerical performance in solving time-dependent PDEs, particularly in long-time prediction of dynamic systems, still needs improvement. In this paper, we focus on solving the long-time integration of nonlinear wave equations via neural operators by replacing the initial condition with the prediction in a recurrent manner. Given limited observed temporal trajectory data, we utilize some intrinsic features of these nonlinear wave equations, such as conservation laws and well-posedness, to improve the algorithm design and reduce accumulated error. Our numerical experiments examine these improvements in the Korteweg-de Vries (KdV) equation, the sine-Gordon equation, and a semilinear wave equation on the irregular domain.

☆ In-Trajectory Inverse Reinforcement Learning: Learn Incrementally From An Ongoing Trajectory

Shicheng Liu, Minghui Zhu

Inverse reinforcement learning (IRL) aims to learn a reward function and a corresponding policy that best fit the demonstrated trajectories of an expert. However, current IRL works cannot learn incrementally from an ongoing trajectory because they have to wait to collect at least one complete trajectory to learn. To bridge the gap, this paper considers the problem of learning a reward function and a corresponding policy while observing the initial state-action pair of an ongoing trajectory and keeping updating the learned reward and policy when new state-action pairs of the ongoing trajectory are observed. We formulate this problem as an online bi-level optimization problem where the upper level dynamically adjusts the learned reward according to the newly observed state-action pairs with the help of a meta-regularization term, and the lower level learns the corresponding policy. We propose a novel algorithm to solve this problem and guarantee that the algorithm achieves sub-linear local regret $O(\sqrt{T}+\log T+\sqrt{T}\log T)$. If the reward function is linear, we prove that the proposed algorithm achieves sub-linear regret $O(\log T)$. Experiments are used to validate the proposed algorithm.

☆ On The Global Convergence Of Online RLHF With Neural Parametrization

Mudit Gaur, Amrit Singh Bedi, Raghu Pasupathy, Vaneet Aggarwal

The importance of Reinforcement Learning from Human Feedback (RLHF) in aligning large language models (LLMs) with human values cannot be overstated. RLHF is a three-stage process that includes supervised fine-tuning (SFT), reward learning, and policy learning. Although there are several offline and online approaches to aligning LLMs, they often suffer from distribution shift issues. These issues arise from the inability to accurately capture the distributional interdependence between the reward learning and policy learning stages. Consequently, this has led to various approximated approaches, but the theoretical insights and motivations remain largely limited to tabular settings, which do not hold in practice. This gap between theoretical insights and practical implementations is critical. It is challenging to address this gap as it requires analyzing the performance of AI alignment algorithms in neural network-parameterized settings. Although bi-level formulations have shown promise in addressing distribution shift issues, they suffer from the hyper-gradient problem, and current approaches lack efficient algorithms to solve this. In this work, we tackle these challenges employing the bi-level formulation laid out in Kwon et al. (2024) along with the assumption \emph{Weak Gradient Domination} to demonstrate convergence in an RLHF setup, obtaining a sample complexity of $\epsilon^{-\frac{7}{2}}$ . Our key contributions are twofold: (i) We propose a bi-level formulation for AI alignment in parameterized settings and introduce a first-order approach to solve this problem. (ii) We analyze the theoretical convergence rates of the proposed algorithm and derive state-of-the-art bounds. To the best of our knowledge, this is the first work to establish convergence rate bounds and global optimality for the RLHF framework in neural network-parameterized settings.

☆ Moonshine: Speech Recognition for Live Transcription and Voice Commands

Nat Jeffries, Evan King, Manjunath Kudlur, Guy Nicholson, James Wang, Pete Warden

This paper introduces Moonshine, a family of speech recognition models optimized for live transcription and voice command processing. Moonshine is based on an encoder-decoder transformer architecture and employs Rotary Position Embedding (RoPE) instead of traditional absolute position embeddings. The model is trained on speech segments of various lengths, but without using zero-padding, leading to greater efficiency for the encoder during inference time. When benchmarked against OpenAI's Whisper tiny.en, Moonshine Tiny demonstrates a 5x reduction in compute requirements for transcribing a 10-second speech segment while incurring no increase in word error rates across standard evaluation datasets. These results highlight Moonshine's potential for real-time and resource-constrained applications.

comment: 7 pages, 6 figures, 3 tables

☆ All You Need is an Improving Column: Enhancing Column Generation for Parallel Machine Scheduling via Transformers

Amira Hijazi, Osman Ozaltin, Reha Uzsoy

We present a neural network-enhanced column generation (CG) approach for a parallel machine scheduling problem. The proposed approach utilizes an encoder-decoder attention model, namely the transformer and pointer architectures, to develop job sequences with negative reduced cost and thus generate columns to add to the master problem. By training the neural network offline and using it in inference mode to predict negative reduced costs columns, we achieve significant computational time savings compared to dynamic programming (DP). Since the exact DP procedure is used to verify that no further columns with negative reduced cost can be identified at termination, the optimality guarantee of the original CG procedure is preserved. For small to medium-sized instances, our approach achieves an average 45% reduction in computation time compared to solving the subproblems with DP. Furthermore, the model generalizes not only to unseen, larger problem instances from the same probability distribution but also to instances from different probability distributions than those presented at training time. For large-sized instances, the proposed approach achieves an 80% improvement in the objective value in under 500 seconds, demonstrating both its scalability and efficiency.

☆ A Comprehensive Comparative Study of Individual ML Models and Ensemble Strategies for Network Intrusion Detection Systems

Ismail Bibers, Osvaldo Arreche, Mustafa Abdallah

The escalating frequency of intrusions in networked systems has spurred the exploration of new research avenues in devising artificial intelligence (AI) techniques for intrusion detection systems (IDS). Various AI techniques have been used to automate network intrusion detection tasks, yet each model possesses distinct strengths and weaknesses. Selecting the optimal model for a given dataset can pose a challenge, necessitating the exploration of ensemble methods to enhance generalization and applicability in network intrusion detection. This paper addresses this gap by conducting a comprehensive evaluation of diverse individual models and both simple and advanced ensemble methods for network IDS. We introduce an ensemble learning framework tailored for assessing individual models and ensemble methods in network intrusion detection tasks. Our framework encompasses the loading of input datasets, training of individual models and ensemble methods, and the generation of evaluation metrics. Furthermore, we incorporate all features across individual models and ensemble techniques. The study presents results for our framework, encompassing 14 methods, including various bagging, stacking, blending, and boosting techniques applied to multiple base learners such as decision trees, neural networks, and among others. We evaluate the framework using two distinct network intrusion datasets, RoEduNet-SIMARGL2021 and CICIDS-2017, each possessing unique characteristics. Additionally, we categorize AI models based on their performances on our evaluation metrics and via their confusion matrices. Our assessment demonstrates the efficacy of learning across most setups explored in this study. Furthermore, we contribute to the community by releasing our source codes, providing a foundational ensemble learning framework for network intrusion detection.

☆ A Comprehensive Survey of Datasets, Theories, Variants, and Applications in Direct Preference Optimization

Wenyi Xiao, Zechuan Wang, Leilei Gan, Shuai Zhao, Wanggui He, Luu Anh Tuan, Long Chen, Hao Jiang, Zhou Zhao, Fei Wu

With the rapid advancement of large language models (LLMs), aligning policy models with human preferences has become increasingly critical. Direct Preference Optimization (DPO) has emerged as a promising approach for alignment, acting as an RL-free alternative to Reinforcement Learning from Human Feedback (RLHF). Despite DPO's various advancements and inherent limitations, an in-depth review of these aspects is currently lacking in the literature. In this work, we present a comprehensive review of the challenges and opportunities in DPO, covering theoretical analyses, variants, relevant preference datasets, and applications. Specifically, we categorize recent studies on DPO based on key research questions to provide a thorough understanding of DPO's current landscape. Additionally, we propose several future research directions to offer insights on model alignment for the research community.

☆ CPE-Pro: A Structure-Sensitive Deep Learning Model for Protein Representation and Origin Evaluation

Wenrui Gou, Wenhui Ge, YangTan, Guisheng Fan, Mingchen Li, Huiqun Yu

Protein structures are important for understanding their functions and interactions. Currently, many protein structure prediction methods are enriching the structure database. Discriminating the origin of structures is crucial for distinguishing between experimentally resolved and computationally predicted structures, evaluating the reliability of prediction methods, and guiding downstream biological studies. Building on works in structure prediction, We developed a structure-sensitive supervised deep learning model, Crystal vs Predicted Evaluator for Protein Structure (CPE-Pro), to represent and discriminate the origin of protein structures. CPE-Pro learns the structural information of proteins and captures inter-structural differences to achieve accurate traceability on four data classes, and is expected to be extended to more. Simultaneously, we utilized Foldseek to encode protein structures into "structure-sequence" and trained a protein Structural Sequence Language Model, SSLM. Preliminary experiments demonstrated that, compared to large-scale protein language models pre-trained on vast amounts of amino acid sequences, the "structure-sequences" enable the language model to learn more informative protein features, enhancing and optimizing structural representations. We have provided the code, model weights, and all related materials on https://github.com/GouWenrui/CPE-Pro-main.git.

☆ SSMT: Few-Shot Traffic Forecasting with Single Source Meta-Transfer ICPR 2024

Kishor Kumar Bhaumik, Minha Kim, Fahim Faisal Niloy, Amin Ahsan Ali, Simon S. Woo

Traffic forecasting in Intelligent Transportation Systems (ITS) is vital for intelligent traffic prediction. Yet, ITS often relies on data from traffic sensors or vehicle devices, where certain cities might not have all those smart devices or enabling infrastructures. Also, recent studies have employed meta-learning to generalize spatial-temporal traffic networks, utilizing data from multiple cities for effective traffic forecasting for data-scarce target cities. However, collecting data from multiple cities can be costly and time-consuming. To tackle this challenge, we introduce Single Source Meta-Transfer Learning (SSMT) which relies only on a single source city for traffic prediction. Our method harnesses this transferred knowledge to enable few-shot traffic forecasting, particularly when the target city possesses limited data. Specifically, we use memory-augmented attention to store the heterogeneous spatial knowledge from the source city and selectively recall them for the data-scarce target city. We extend the idea of sinusoidal positional encoding to establish meta-learning tasks by leveraging diverse temporal traffic patterns from the source city. Moreover, to capture a more generalized representation of the positions we introduced a meta-positional encoding that learns the most optimal representation of the temporal pattern across all the tasks. We experiment on five real-world benchmark datasets to demonstrate that our method outperforms several existing methods in time series traffic prediction.

comment: ICPR 2024

☆ Multimodal Learning for Embryo Viability Prediction in Clinical IVF MICCAI 2024

Junsik Kim, Zhiyi Shi, Davin Jeong, Johannes Knittel, Helen Y. Yang, Yonghyun Song, Wanhua Li, Yicong Li, Dalit Ben-Yosef, Daniel Needleman, Hanspeter Pfister

In clinical In-Vitro Fertilization (IVF), identifying the most viable embryo for transfer is important to increasing the likelihood of a successful pregnancy. Traditionally, this process involves embryologists manually assessing embryos' static morphological features at specific intervals using light microscopy. This manual evaluation is not only time-intensive and costly, due to the need for expert analysis, but also inherently subjective, leading to variability in the selection process. To address these challenges, we develop a multimodal model that leverages both time-lapse video data and Electronic Health Records (EHRs) to predict embryo viability. One of the primary challenges of our research is to effectively combine time-lapse video and EHR data, owing to their inherent differences in modality. We comprehensively analyze our multimodal model with various modality inputs and integration approaches. Our approach will enable fast and automated embryo viability predictions in scale for clinical IVF.

comment: Accepted to MICCAI 2024

☆ Language Models are Symbolic Learners in Arithmetic

Chunyuan Deng, Zhiqi Li, Roy Xie, Ruidi Chang, Hanjie Chen

Large Language Models (LLMs) are thought to struggle with arithmetic learning due to the inherent differences between language modeling and numerical computation, but concrete evidence has been lacking. This work responds to this claim through a two-side experiment. We first investigate whether LLMs leverage partial products during arithmetic learning. We find that although LLMs can identify some partial products after learning, they fail to leverage them for arithmetic tasks, conversely. We then explore how LLMs approach arithmetic symbolically by breaking tasks into subgroups, hypothesizing that difficulties arise from subgroup complexity and selection. Our results show that when subgroup complexity is fixed, LLMs treat a collection of different arithmetic operations similarly. By analyzing position-level accuracy across different training sizes, we further observe that it follows a U-shaped pattern: LLMs quickly learn the easiest patterns at the first and last positions, while progressively learning the more difficult patterns in the middle positions. This suggests that LLMs select subgroup following an easy-to-hard paradigm during learning. Our work confirms that LLMs are pure symbolic learners in arithmetic tasks and underscores the importance of understanding them deeply through subgroup-level quantification.

☆ Generalized Probabilistic Attention Mechanism in Transformers

DongNyeong Heo, Heeyoul Choi

The Transformer architecture has become widely adopted due to its demonstrated success, attributed to the attention mechanism at its core. Despite these successes, the attention mechanism of Transformers is associated with two well-known issues: rank-collapse and gradient vanishing. In this paper, we present a theoretical analysis that it is inherently difficult to address both issues simultaneously in the conventional attention mechanism. To handle these issues, we introduce a novel class of attention mechanism, referred to as generalized probabilistic attention mechanism (GPAM), and its dual-attention implementation within the Transformer architecture. Unlike conventional attention mechanisms, GPAM allows for negative attention scores while preserving a fixed total sum. We provide theoretical evidence that the proposed dual-attention GPAM (daGPAM) effectively mitigates both the rank-collapse and gradient vanishing issues which are difficult to resolve simultaneously with the conventional attention mechanisms. Furthermore, we empirically validate this theoretical evidence, demonstrating the superiority of daGPAM compared to other alternative attention mechanisms that were proposed to address the same issues. Additionally, we demonstrate the practical benefits of GPAM in natural language processing tasks, such as language modeling and neural machine translation.

☆ Stacking Small Language Models for Generalizability

Laurence Liang

Recent advances show that large language models (LLMs) generalize strong performance across different natural language benchmarks. However, the large size of LLMs makes training and inference expensive and impractical to run in resource-limited settings. This paper introduces a new approach called fine-tuning stacks of language models (FSLM), which involves stacking small language models (SLM) as an alternative to LLMs. By fine-tuning each SLM to perform a specific task, this approach breaks down high level reasoning into multiple lower-level steps that specific SLMs are responsible for. As a result, FSLM allows for lower training and inference costs, and also improves model interpretability as each SLM communicates with the subsequent one through natural language. By evaluating FSLM on common natural language benchmarks, this paper highlights promising early results toward generalizable performance using FSLM as a cost-effective alternative to LLMs.

☆ Pruning Foundation Models for High Accuracy without Retraining EMNLP 2024

Pu Zhao, Fei Sun, Xuan Shen, Pinrui Yu, Zhenglun Kong, Yanzhi Wang, Xue Lin

Despite the superior performance, it is challenging to deploy foundation models or large language models (LLMs) due to their massive parameters and computations. While pruning is a promising technique to reduce model size and accelerate the inference, the traditional pruning techniques can hardly be applied for LLMs as they need to finetune the model on the full dataset with multiple epochs consuming massive data and hardware resources. To deal with this problem, post-training pruning methods are proposed to prune LLMs in one-shot without retraining. However, their accuracy after pruning may suffer from certain performance degradation due to the lack of retraining with massive data. To address this issue, in this paper, we first formulate the post-training problem for layer-wise LLM compression to simultaneously prune multiple weights in LLMs. Next, we provide an optimal solution for this problem and design our post-training pruning algorithm for both unstructured and semi-structured sparsity. Our extensive experiments demonstrate the superior performance of the proposed methods in comparison to SOTA baselines across various LLM families including transformer-based LLMs and Mamba-based LLMs. Code link: https://github.com/piuzha/APT

comment: Accepted by EMNLP 2024 findings

☆ Reward Maximization for Pure Exploration: Minimax Optimal Good Arm Identification for Nonparametric Multi-Armed Bandits

Brian Cho, Dominik Meier, Kyra Gan, Nathan Kallus

In multi-armed bandits, the tasks of reward maximization and pure exploration are often at odds with each other. The former focuses on exploiting arms with the highest means, while the latter may require constant exploration across all arms. In this work, we focus on good arm identification (GAI), a practical bandit inference objective that aims to label arms with means above a threshold as quickly as possible. We show that GAI can be efficiently solved by combining a reward-maximizing sampling algorithm with a novel nonparametric anytime-valid sequential test for labeling arm means. We first establish that our sequential test maintains error control under highly nonparametric assumptions and asymptotically achieves the minimax optimal e-power, a notion of power for anytime-valid tests. Next, by pairing regret-minimizing sampling schemes with our sequential test, we provide an approach that achieves minimax optimal stopping times for labeling arms with means above a threshold, under an error probability constraint. Our empirical results validate our approach beyond the minimax setting, reducing the expected number of samples for all stopping times by at least 50% across both synthetic and real-world settings.

☆ How to Find the Exact Pareto Front for Multi-Objective MDPs?

Yining Li, Peizhong Ju, Ness B. Shroff

Multi-objective Markov Decision Processes (MDPs) are receiving increasing attention, as real-world decision-making problems often involve conflicting objectives that cannot be addressed by a single-objective MDP. The Pareto front identifies the set of policies that cannot be dominated, providing a foundation for finding optimal solutions that can efficiently adapt to various preferences. However, finding the Pareto front is a highly challenging problem. Most existing methods either (i) rely on traversing the continuous preference space, which is impractical and results in approximations that are difficult to evaluate against the true Pareto front, or (ii) focus solely on deterministic Pareto optimal policies, from which there are no known techniques to characterize the full Pareto front. Moreover, finding the structure of the Pareto front itself remains unclear even in the context of dynamic programming. This work addresses the challenge of efficiently discovering the Pareto front. By investigating the geometric structure of the Pareto front in MO-MDP, we uncover a key property: the Pareto front is on the boundary of a convex polytope whose vertices all correspond to deterministic policies, and neighboring vertices of the Pareto front differ by only one state-action pair of the deterministic policy, almost surely. This insight transforms the global comparison across all policies into a localized search among deterministic policies that differ by only one state-action pair, drastically reducing the complexity of searching for the exact Pareto front. We develop an efficient algorithm that identifies the vertices of the Pareto front by solving a single-objective MDP only once and then traversing the edges of the Pareto front, making it more efficient than existing methods. Our empirical studies demonstrate the effectiveness of our theoretical strategy in discovering the Pareto front.

☆ Gradient Rewiring for Editable Graph Neural Network Training NeurIPS 2024

Zhimeng Jiang, Zirui Liu, Xiaotian Han, Qizhang Feng, Hongye Jin, Qiaoyu Tan, Kaixiong Zhou, Na Zou, Xia Hu

Deep neural networks are ubiquitously adopted in many applications, such as computer vision, natural language processing, and graph analytics. However, well-trained neural networks can make prediction errors after deployment as the world changes. \textit{Model editing} involves updating the base model to correct prediction errors with less accessible training data and computational resources. Despite recent advances in model editors in computer vision and natural language processing, editable training in graph neural networks (GNNs) is rarely explored. The challenge with editable GNN training lies in the inherent information aggregation across neighbors, which can lead model editors to affect the predictions of other nodes unintentionally. In this paper, we first observe the gradient of cross-entropy loss for the target node and training nodes with significant inconsistency, which indicates that directly fine-tuning the base model using the loss on the target node deteriorates the performance on training nodes. Motivated by the gradient inconsistency observation, we propose a simple yet effective \underline{G}radient \underline{R}ewiring method for \underline{E}ditable graph neural network training, named \textbf{GRE}. Specifically, we first store the anchor gradient of the loss on training nodes to preserve the locality. Subsequently, we rewire the gradient of the loss on the target node to preserve performance on the training node using anchor gradient. Experiments demonstrate the effectiveness of GRE on various model architectures and graph datasets in terms of multiple editing situations. The source code is available at \url{https://github.com/zhimengj0326/Gradient_rewiring_editing}

comment: NeurIPS 2024

☆ Bayesian Concept Bottleneck Models with LLM Priors

Jean Feng, Avni Kothari, Luke Zier, Chandan Singh, Yan Shuo Tan

Concept Bottleneck Models (CBMs) have been proposed as a compromise between white-box and black-box models, aiming to achieve interpretability without sacrificing accuracy. The standard training procedure for CBMs is to predefine a candidate set of human-interpretable concepts, extract their values from the training data, and identify a sparse subset as inputs to a transparent prediction model. However, such approaches are often hampered by the tradeoff between enumerating a sufficiently large set of concepts to include those that are truly relevant versus controlling the cost of obtaining concept extractions. This work investigates a novel approach that sidesteps these challenges: BC-LLM iteratively searches over a potentially infinite set of concepts within a Bayesian framework, in which Large Language Models (LLMs) serve as both a concept extraction mechanism and prior. BC-LLM is broadly applicable and multi-modal. Despite imperfections in LLMs, we prove that BC-LLM can provide rigorous statistical inference and uncertainty quantification. In experiments, it outperforms comparator methods including black-box models, converges more rapidly towards relevant concepts and away from spuriously correlated ones, and is more robust to out-of-distribution samples.

☆ A Plug-and-Play Fully On-the-Job Real-Time Reinforcement Learning Algorithm for a Direct-Drive Tandem-Wing Experiment Platforms Under Multiple Random Operating Conditions

Zhang Minghao, Song Bifeng, Yang Xiaojun, Wang Liang

The nonlinear and unstable aerodynamic interference generated by the tandem wings of such biomimetic systems poses substantial challenges for motion control, especially under multiple random operating conditions. To address these challenges, the Concerto Reinforcement Learning Extension (CRL2E) algorithm has been developed. This plug-and-play, fully on-the-job, real-time reinforcement learning algorithm incorporates a novel Physics-Inspired Rule-Based Policy Composer Strategy with a Perturbation Module alongside a lightweight network optimized for real-time control. To validate the performance and the rationality of the module design, experiments were conducted under six challenging operating conditions, comparing seven different algorithms. The results demonstrate that the CRL2E algorithm achieves safe and stable training within the first 500 steps, improving tracking accuracy by 14 to 66 times compared to the Soft Actor-Critic, Proximal Policy Optimization, and Twin Delayed Deep Deterministic Policy Gradient algorithms. Additionally, CRL2E significantly enhances performance under various random operating conditions, with improvements in tracking accuracy ranging from 8.3% to 60.4% compared to the Concerto Reinforcement Learning (CRL) algorithm. The convergence speed of CRL2E is 36.11% to 57.64% faster than the CRL algorithm with only the Composer Perturbation and 43.52% to 65.85% faster than the CRL algorithm when both the Composer Perturbation and Time-Interleaved Capability Perturbation are introduced, especially in conditions where the standard CRL struggles to converge. Hardware tests indicate that the optimized lightweight network structure excels in weight loading and average inference time, meeting real-time control requirements.

comment: 63 pages, 32 figures

☆ Hiding in Plain Sight: Reframing Hardware Trojan Benchmarking as a Hide&Seek Modification

Amin Sarihi, Ahmad Patooghy, Peter Jamieson, Abdel-Hameed A. Badawy

This work focuses on advancing security research in the hardware design space by formally defining the realistic problem of Hardware Trojan (HT) detection. The goal is to model HT detection more closely to the real world, i.e., describing the problem as The Seeker's Dilemma where a detecting agent is unaware of whether circuits are infected by HTs or not. Using this theoretical problem formulation, we create a benchmark that consists of a mixture of HT-free and HT-infected restructured circuits while preserving their original functionalities. The restructured circuits are randomly infected by HTs, causing a situation where the defender is uncertain if a circuit is infected or not. We believe that our innovative benchmark and methodology of creating benchmarks will help the community judge the detection quality of different methods by comparing their success rates in circuit classification. We use our developed benchmark to evaluate three state-of-the-art HT detection tools to show baseline results for this approach. We use Principal Component Analysis to assess the strength of our benchmark, where we observe that some restructured HT-infected circuits are mapped closely to HT-free circuits, leading to significant label misclassification by detectors.

comment: arXiv admin note: substantial text overlap with arXiv:2402.17918

☆ Distributed Thompson sampling under constrained communication

Saba Zerefa, Zhaolin Ren, Haitong Ma, Na Li

In Bayesian optimization, a black-box function is maximized via the use of a surrogate model. We apply distributed Thompson sampling, using a Gaussian process as a surrogate model, to approach the multi-agent Bayesian optimization problem. In our distributed Thompson sampling implementation, each agent receives sampled points from neighbors, where the communication network is encoded in a graph; each agent utilizes a Gaussian process to model the objective function. We demonstrate a theoretical bound on Bayesian Simple Regret, where the bound depends on the size of the largest complete subgraph of the communication graph. Unlike in batch Bayesian optimization, this bound is applicable in cases where the communication graph amongst agents is constrained. When compared to sequential Thompson sampling, our bound guarantees faster convergence with respect to time as long as there is a fully connected subgraph of at least two agents. We confirm the efficacy of our algorithm with numerical simulations on traditional optimization test functions, illustrating the significance of graph connectivity on improving regret convergence.

comment: 9 pages

♻ ☆ Dual-Label Learning With Irregularly Present Labels

Mingqian Li, Qiao Han, Yiteng Zhai, Ruifeng Li, Yao Yang, Hongyang Chen

In multi-task learning, we often encounter the case when the presence of labels across samples exhibits irregular patterns: samples can be fully labeled, partially labeled or unlabeled. Taking drug analysis as an example, multiple toxicity properties of a drug molecule may not be concurrently available due to experimental limitations. It triggers a demand for a new training and inference mechanism that could accommodate irregularly present labels and maximize the utility of any available label information. In this work, we focus on the two-label learning task, and propose a novel training and inference framework, Dual-Label Learning (DLL). The DLL framework formulates the problem into a dual-function system, in which the two functions should simultaneously satisfy standard supervision, structural duality and probabilistic duality. DLL features a dual-tower model architecture that explicitly captures the information exchange between labels, aimed at maximizing the utility of partially available labels in understanding label correlation. During training, label imputation for missing labels is conducted as part of the forward propagation process, while during inference, labels are regarded as unknowns of a bivariate system of equations and are solved jointly. Theoretical analysis guarantees the feasibility of DLL, and extensive experiments are conducted to verify that by explicitly modeling label correlation and maximizing the utility of available labels, our method makes consistently better predictions than baseline approaches by up to a 10% gain in F1-score or MAPE. Remarkably, our method provided with data at a label missing rate as high as 60% can achieve similar or even better results than baseline approaches at a label missing rate of only 10%.

♻ ☆ RILe: Reinforced Imitation Learning

Mert Albaba, Sammy Christen, Thomas Langarek, Christoph Gebhardt, Otmar Hilliges, Michael J. Black

Reinforcement Learning has achieved significant success in generating complex behavior but often requires extensive reward function engineering. Adversarial variants of Imitation Learning and Inverse Reinforcement Learning offer an alternative by learning policies from expert demonstrations via a discriminator. However, these methods struggle in complex tasks where randomly sampling expert-like behaviors is challenging. This limitation stems from their reliance on policy-agnostic discriminators, which provide insufficient guidance for agent improvement, especially as task complexity increases and expert behavior becomes more distinct. We introduce RILe (Reinforced Imitation Learning environment), a novel trainer-student system that learns a dynamic reward function based on the student's performance and alignment with expert demonstrations. In RILe, the student learns an action policy while the trainer, using reinforcement learning, continuously updates itself via the discriminator's feedback to optimize the alignment between the student and the expert. The trainer optimizes for long-term cumulative rewards from the discriminator, enabling it to provide nuanced feedback that accounts for the complexity of the task and the student's current capabilities. This approach allows for greater exploration of agent actions by providing graduated feedback rather than binary expert/non-expert classifications. By reducing dependence on policy-agnostic discriminators, RILe enables better performance in complex settings where traditional methods falter, outperforming existing methods by 2x in complex simulated robot-locomotion tasks.

♻ ☆ Comparing the information content of probabilistic representation spaces

Kieran A. Murphy, Sam Dillavou, Dani S. Bassett

Probabilistic representation spaces convey information about a dataset, and to understand the effects of factors such as training loss and network architecture, we seek to compare the information content of such spaces. However, most existing methods to compare representation spaces assume representations are points, and neglect the distributional nature of probabilistic representations. Here, instead of building upon point-based measures of comparison, we build upon classic methods from literature on hard clustering. We generalize two information-theoretic methods of comparing hard clustering assignments to be applicable to general probabilistic representation spaces. We then propose a practical method of estimation that is based on fingerprinting a representation space with a sample of the dataset and is applicable when the communicated information is only a handful of bits. With unsupervised disentanglement as a motivating problem, we find information fragments that are repeatedly contained in individual latent dimensions in VAE and InfoGAN ensembles. Then, by comparing the full latent spaces of models, we find highly consistent information content across datasets, methods, and hyperparameters, even though there is often a point during training with substantial variety across repeat runs. Finally, we leverage the differentiability of the proposed method and perform model fusion by synthesizing the information content of multiple weak learners, each incapable of representing the global structure of a dataset. Across the case studies, the direct comparison of information content provides a natural basis for understanding the processing of information.

comment: Code: https://github.com/murphyka/representation-space-info-comparison

♻ ☆ Hypergraph: A Unified and Uniform Definition with Application to Chemical Hypergraph and More

Daniel T. Chang

The conventional definition of hypergraph has two major issues: (1) there is not a standard definition of directed hypergraph and (2) there is not a formal definition of nested hypergraph. To resolve these issues, we propose a new definition of hypergraph that unifies the concepts of undirected, directed and nested hypergraphs, and that is uniform in using hyperedge as a single construct for representing high-order correlations among things, i.e., nodes and hyperedges. Specifically, we define a hyperedge to be a simple hyperedge, a nesting hyperedge, or a directed hyperedge. With this new definition, a hypergraph is nested if it has nesting hyperedge(s), and is directed if it has directed hyperedge(s). Otherwise, a hypergraph is a simple hypergraph. The uniformity and power of this new definition, with visualization, should facilitate the use of hypergraph for representing (hierarchical) high-order correlations in general and chemical systems in particular. Graph has been widely used as a mathematical structure for machine learning on molecular structures and 3D molecular geometries. However, graph has a major limitation: it can represent only pairwise correlations between nodes. Hypergraph extends graph with high-order correlations among nodes. This extension is significant or essential for machine learning on chemical systems. For molecules, this is significant as it allows the direct, explicit representation of multicenter bonds and molecular substructures. For chemical reactions, this is essential since most chemical reactions involve multiple participants. We propose the use of chemical hypergraph, a multilevel hypergraph with simple, nesting and directed hyperedges, as a single mathematical structure for representing chemical systems. We apply the new definition of hypergraph to chemical hypergraph and, as simplified versions, molecular hypergraph and chemical reaction hypergraph.

comment: arXiv admin note: text overlap with arXiv:2310.03623 by other authors

♻ ☆ Bidirectional Decoding: Improving Action Chunking via Closed-Loop Resampling

Yuejiang Liu, Jubayer Ibn Hamid, Annie Xie, Yoonho Lee, Maximilian Du, Chelsea Finn

Predicting and executing a sequence of actions without intermediate replanning, known as action chunking, is increasingly used in robot learning from human demonstrations. Yet, its reported effects on the learned policy are inconsistent: some studies find it crucial for achieving strong results, while others observe decreased performance. In this paper, we first dissect how action chunking impacts the divergence between a learner and a demonstrator. We find that action chunking allows the learner to better capture the temporal dependencies in demonstrations but at the cost of reduced reactivity in stochastic environments. To address this tradeoff, we propose Bidirectional Decoding (BID), a test-time inference algorithm that bridges action chunking with closed-loop operations. BID samples multiple predictions at each time step and searches for the optimal one based on two criteria: (i) backward coherence, which favors samples that align with previous decisions; (ii) forward contrast, which seeks samples of high likelihood for future plans. By coupling decisions within and across action chunks, BID promotes consistency over time while maintaining reactivity to unexpected changes. Experimental results show that BID boosts the performance of two state-of-the-art generative policies across seven simulation benchmarks and two real-world tasks. Code and videos are available at https://bid-robot.github.io.

comment: Project website: https://bid-robot.github.io/

♻ ☆ Decomposing and Interpreting Image Representations via Text in ViTs Beyond CLIP NeurIPS 2024

Sriram Balasubramanian, Samyadeep Basu, Soheil Feizi

Recent work has explored how individual components of the CLIP-ViT model contribute to the final representation by leveraging the shared image-text representation space of CLIP. These components, such as attention heads and MLPs, have been shown to capture distinct image features like shape, color or texture. However, understanding the role of these components in arbitrary vision transformers (ViTs) is challenging. To this end, we introduce a general framework which can identify the roles of various components in ViTs beyond CLIP. Specifically, we (a) automate the decomposition of the final representation into contributions from different model components, and (b) linearly map these contributions to CLIP space to interpret them via text. Additionally, we introduce a novel scoring function to rank components by their importance with respect to specific features. Applying our framework to various ViT variants (e.g. DeiT, DINO, DINOv2, Swin, MaxViT), we gain insights into the roles of different components concerning particular image features. These insights facilitate applications such as image retrieval using text descriptions or reference images, visualizing token importance heatmaps, and mitigating spurious correlations. We release our code to reproduce the experiments at https://github.com/SriramB-98/vit-decompose

comment: NeurIPS 2024, 31 pages, 15 figures

♻ ☆ Generation through the lens of learning theory

Vinod Raman, Ambuj Tewari

We study generation through the lens of statistical learning theory. First, we abstract and formalize the results of Gold [1967], Angluin [1979, 1980], and Kleinberg and Mullainathan [2024] for language identification/generation in the limit in terms of a binary hypothesis class defined over an abstract instance space. Then, we formalize a different paradigm of generation studied by Kleinberg and Mullainathan [2024], which we call ``uniform generation," and provide a characterization of which hypothesis classes are uniformly generatable. As is standard in statistical learning theory, our characterization is in terms of the finiteness of a new combinatorial dimension we call the Closure dimension. By doing so, we are able to compare generatability with predictability (captured via PAC and online learnability) and show that these two properties of hypothesis classes are \emph{incompatible} - there are classes that are generatable but not predictable and vice versa.

comment: Minor edits

♻ ☆ Are Large Language Models Chameleons? An Attempt to Simulate Social Surveys

Mingmeng Geng, Sihong He, Roberto Trotta

Can large language models (LLMs) simulate social surveys? To answer this question, we conducted millions of simulations in which LLMs were asked to answer subjective questions. A comparison of different LLM responses with the European Social Survey (ESS) data suggests that the effect of prompts on bias and variability is fundamental, highlighting major cultural, age, and gender biases. We further discussed statistical methods for measuring the difference between LLM answers and survey data and proposed a novel measure inspired by Jaccard similarity, as LLM-generated responses are likely to have a smaller variance. Our experiments also reveal that it is important to analyze the robustness and variability of prompts before using LLMs to simulate social surveys, as their imitation abilities are approximate at best.

comment: 17 pages

♻ ☆ IoT-Based Preventive Mental Health Using Knowledge Graphs and Standards for Better Well-Being

Amelie Gyrard, Seyedali Mohammadi, Manas Gaur, Antonio Kung

Sustainable Development Goals (SDGs) give the UN a road map for development with Agenda 2030 as a target. SDG3 "Good Health and Well-Being" ensures healthy lives and promotes well-being for all ages. Digital technologies can support SDG3. Burnout and even depression could be reduced by encouraging better preventive health. Due to the lack of patient knowledge and focus to take care of their health, it is necessary to help patients before it is too late. New trends such as positive psychology and mindfulness are highly encouraged in the USA. Digital Twins (DTs) can help with the continuous monitoring of emotion using physiological signals (e.g., collected via wearables). DTs facilitate monitoring and provide constant health insight to improve quality of life and well-being with better personalization. Healthcare DTs challenges are standardizing data formats, communication protocols, and data exchange mechanisms. As an example, ISO has the ISO/IEC JTC 1/SC 41 Internet of Things (IoT) and DTs Working Group, with standards such as "ISO/IEC 21823-3:2021 IoT - Interoperability for IoT Systems - Part 3 Semantic interoperability", "ISO/IEC CD 30178 - IoT - Data format, value and coding". To achieve those data integration and knowledge challenges, we designed the Mental Health Knowledge Graph (ontology and dataset) to boost mental health. As an example, explicit knowledge is described such as chocolate contains magnesium which is recommended for depression. The Knowledge Graph (KG) acquires knowledge from ontology-based mental health projects classified within the LOV4IoT ontology catalog (Emotion, Depression, and Mental Health). Furthermore, the KG is mapped to standards when possible. Standards from ETSI SmartM2M can be used such as SAREF4EHAW to represent medical devices and sensors, but also ITU/WHO, ISO, W3C, NIST, and IEEE standards relevant to mental health can be considered.

comment: 20 pages, Book chapter, Smart Technologies for Achieving Good Health and Well-Being: Towards Sustainable Development Goal, Taylor & Francis

♻ ☆ Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey

Tiansheng Huang, Sihao Hu, Fatih Ilhan, Selim Furkan Tekin, Ling Liu

Recent research demonstrates that the nascent fine-tuning-as-a-service business model exposes serious safety concerns -- fine-tuning over a few harmful data uploaded by the users can compromise the safety alignment of the model. The attack, known as harmful fine-tuning, has raised a broad research interest among the community. However, as the attack is still new, \textbf{we observe from our miserable submission experience that there are general misunderstandings within the research community.} We in this paper aim to clear some common concerns for the attack setting, and formally establish the research problem. Specifically, we first present the threat model of the problem, and introduce the harmful fine-tuning attack and its variants. Then we systematically survey the existing literature on attacks/defenses/mechanical analysis of the problem. Finally, we outline future research directions that might contribute to the development of the field. Additionally, we present a list of questions of interest, which might be useful to refer to when reviewers in the peer review process question the realism of the experiment/attack/defense setting. A curated list of relevant papers is maintained and made accessible at: \url{https://github.com/git-disl/awesome_LLM-harmful-fine-tuning-papers}.

♻ ☆ SleeperNets: Universal Backdoor Poisoning Attacks Against Reinforcement Learning Agents NeurIPS

Ethan Rathbun, Christopher Amato, Alina Oprea

Reinforcement learning (RL) is an actively growing field that is seeing increased usage in real-world, safety-critical applications -- making it paramount to ensure the robustness of RL algorithms against adversarial attacks. In this work we explore a particularly stealthy form of training-time attacks against RL -- backdoor poisoning. Here the adversary intercepts the training of an RL agent with the goal of reliably inducing a particular action when the agent observes a pre-determined trigger at inference time. We uncover theoretical limitations of prior work by proving their inability to generalize across domains and MDPs. Motivated by this, we formulate a novel poisoning attack framework which interlinks the adversary's objectives with those of finding an optimal policy -- guaranteeing attack success in the limit. Using insights from our theoretical analysis we develop ``SleeperNets'' as a universal backdoor attack which exploits a newly proposed threat model and leverages dynamic reward poisoning techniques. We evaluate our attack in 6 environments spanning multiple domains and demonstrate significant improvements in attack success over existing methods, while preserving benign episodic return.

comment: 23 pages, 14 figures, NeurIPS

♻ ☆ This Too Shall Pass: Removing Stale Observations in Dynamic Bayesian Optimization

Anthony Bardou, Patrick Thiran, Giovanni Ranieri

Bayesian Optimization (BO) has proven to be very successful at optimizing a static, noisy, costly-to-evaluate black-box function $f : \mathcal{S} \to \mathbb{R}$. However, optimizing a black-box which is also a function of time (i.e., a dynamic function) $f : \mathcal{S} \times \mathcal{T} \to \mathbb{R}$ remains a challenge, since a dynamic Bayesian Optimization (DBO) algorithm has to keep track of the optimum over time. This changes the nature of the optimization problem in at least three aspects: (i) querying an arbitrary point in $\mathcal{S} \times \mathcal{T}$ is impossible, (ii) past observations become less and less relevant for keeping track of the optimum as time goes by and (iii) the DBO algorithm must have a high sampling frequency so it can collect enough relevant observations to keep track of the optimum through time. In this paper, we design a Wasserstein distance-based criterion able to quantify the relevancy of an observation with respect to future predictions. Then, we leverage this criterion to build W-DBO, a DBO algorithm able to remove irrelevant observations from its dataset on the fly, thus maintaining simultaneously a good predictive performance and a high sampling frequency, even in continuous-time optimization tasks with unknown horizon. Numerical experiments establish the superiority of W-DBO, which outperforms state-of-the-art methods by a comfortable margin.

♻ ☆ Differentiable Optimization of Similarity Scores Between Models and Brains

Nathan Cloos, Moufan Li, Markus Siegel, Scott L. Brincat, Earl K. Miller, Guangyu Robert Yang, Christopher J. Cueva

How do we know if two systems - biological or artificial - process information in a similar way? Similarity measures such as linear regression, Centered Kernel Alignment (CKA), Normalized Bures Similarity (NBS), and angular Procrustes distance, are often used to quantify this similarity. However, it is currently unclear what drives high similarity scores and even what constitutes a "good" score. Here, we introduce a novel tool to investigate these questions by differentiating through similarity measures to directly maximize the score. Surprisingly, we find that high similarity scores do not guarantee encoding task-relevant information in a manner consistent with neural data; and this is particularly acute for CKA and even some variations of cross-validated and regularized linear regression. We find no consistent threshold for a good similarity score - it depends on both the measure and the dataset. In addition, synthetic datasets optimized to maximize similarity scores initially learn the highest variance principal component of the target dataset, but some methods like angular Procrustes capture lower variance dimensions much earlier than methods like CKA. To shed light on this, we mathematically derive the sensitivity of CKA, angular Procrustes, and NBS to the variance of principal component dimensions, and explain the emphasis CKA places on high variance components. Finally, by jointly optimizing multiple similarity measures, we characterize their allowable ranges and reveal that some similarity measures are more constraining than others. While current measures offer a seemingly straightforward way to quantify the similarity between neural systems, our work underscores the need for careful interpretation. We hope the tools we developed will be used by practitioners to better understand current and future similarity measures.

comment: 19 pages, 9 figures

♻ ☆ Adaptive $Q$-Network: On-the-fly Target Selection for Deep Reinforcement Learning

Théo Vincent, Fabian Wahren, Jan Peters, Boris Belousov, Carlo D'Eramo

Deep Reinforcement Learning (RL) is well known for being highly sensitive to hyperparameters, requiring practitioners substantial efforts to optimize them for the problem at hand. This also limits the applicability of RL in real-world scenarios. In recent years, the field of automated Reinforcement Learning (AutoRL) has grown in popularity by trying to address this issue. However, these approaches typically hinge on additional samples to select well-performing hyperparameters, hindering sample-efficiency and practicality. Furthermore, most AutoRL methods are heavily based on already existing AutoML methods, which were originally developed neglecting the additional challenges inherent to RL due to its non-stationarities. In this work, we propose a new approach for AutoRL, called Adaptive $Q$-Network (AdaQN), that is tailored to RL to take into account the non-stationarity of the optimization procedure without requiring additional samples. AdaQN learns several $Q$-functions, each one trained with different hyperparameters, which are updated online using the $Q$-function with the smallest approximation error as a shared target. Our selection scheme simultaneously handles different hyperparameters while coping with the non-stationarity induced by the RL optimization procedure and being orthogonal to any critic-based RL algorithm. We demonstrate that AdaQN is theoretically sound and empirically validate it in MuJoCo control problems and Atari $2600$ games, showing benefits in sample-efficiency, overall performance, robustness to stochasticity and training stability.

comment: Preprint

♻ ☆ Adversarial Inception for Bounded Backdoor Poisoning in Deep Reinforcement Learning ICLR 2025

Ethan Rathbun, Christopher Amato, Alina Oprea

Recent works have demonstrated the vulnerability of Deep Reinforcement Learning (DRL) algorithms against training-time, backdoor poisoning attacks. These attacks induce pre-determined, adversarial behavior in the agent upon observing a fixed trigger during deployment while allowing the agent to solve its intended task during training. Prior attacks rely on arbitrarily large perturbations to the agent's rewards to achieve both of these objectives - leaving them open to detection. Thus, in this work, we propose a new class of backdoor attacks against DRL which achieve state of the art performance while minimally altering the agent's rewards. These "inception" attacks train the agent to associate the targeted adversarial behavior with high returns by inducing a disjunction between the agent's chosen action and the true action executed in the environment during training. We formally define these attacks and prove they can achieve both adversarial objectives. We then devise an online inception attack which significantly out-performs prior attacks under bounded reward constraints.

comment: 10 pages, 5 figures, ICLR 2025

♻ ☆ Exploring the Potential of Large Language Models for Heterophilic Graphs

Yuxia Wu, Shujie Li, Yuan Fang, Chuan Shi

Large language models (LLMs) have presented significant opportunities to enhance various machine learning applications, including graph neural networks (GNNs). By leveraging the vast open-world knowledge within LLMs, we can more effectively interpret and utilize textual data to better characterize heterophilic graphs, where neighboring nodes often have different labels. However, existing approaches for heterophilic graphs overlook the rich textual data associated with nodes, which could unlock deeper insights into their heterophilic contexts. In this work, we explore the potential of LLMs for modeling heterophilic graphs and propose a novel two-stage framework: LLM-enhanced edge discriminator and LLM-guided edge reweighting. In the first stage, we fine-tune the LLM to better identify homophilic and heterophilic edges based on the textual content of their nodes. In the second stage, we adaptively manage message propagation in GNNs for different edge types based on node features, structures, and heterophilic or homophilic characteristics. To cope with the computational demands when deploying LLMs in practical scenarios, we further explore model distillation techniques to fine-tune smaller, more efficient models that maintain competitive performance. Extensive experiments validate the effectiveness of our framework, demonstrating the feasibility of using LLMs to enhance node classification on heterophilic graphs.

comment: Under review

♻ ☆ MathGAP: Out-of-Distribution Evaluation on Problems with Arbitrarily Complex Proofs

Andreas Opedal, Haruki Shirakami, Bernhard Schölkopf, Abulhair Saparov, Mrinmaya Sachan

Large language models (LLMs) can solve arithmetic word problems with high accuracy, but little is known about how well they generalize to problems that are more complex than the ones on which they have been trained. Empirical investigations of such questions are impeded by two major flaws of current evaluations: (i) much of the evaluation data is contaminated, in the sense that it has already been seen during training, and (ii) benchmark datasets do not capture how problem proofs may be arbitrarily complex in various ways. As a step towards addressing these issues, we present a framework for evaluating LLMs on problems with arbitrarily complex arithmetic proofs, called MathGAP. MathGAP generates problems that follow fixed proof specifications -- along with chain-of-thought reasoning annotations -- enabling systematic studies on generalization with respect to arithmetic proof complexity. We apply MathGAP to analyze how in-context learning interacts with generalization to problems that have more complex proofs. We find that among the models tested, most show a significant decrease in performance as proofs get deeper and wider. This effect is more pronounced in complex, nonlinear proof structures, which are challenging even for GPT-4o. Surprisingly, providing in-context examples from the same distribution as the test set is not always beneficial for performance. In particular, zero-shot prompting as well as demonstrating a diverse range of examples that are less complex than the test data sometimes yield similar or higher accuracies.

comment: Preprint

♻ ☆ Human-Agent Joint Learning for Efficient Robot Manipulation Skill Acquisition

Shengcheng Luo, Quanquan Peng, Jun Lv, Kaiwen Hong, Katherine Rose Driggs-Campbell, Cewu Lu, Yong-Lu Li

Employing a teleoperation system for gathering demonstrations offers the potential for more efficient learning of robot manipulation. However, teleoperating a robot arm equipped with a dexterous hand or gripper, via a teleoperation system presents inherent challenges due to the task's high dimensionality, complexity of motion, and differences between physiological structures. In this study, we introduce a novel system for joint learning between human operators and robots, that enables human operators to share control of a robot end-effector with a learned assistive agent, simplifies the data collection process, and facilitates simultaneous human demonstration collection and robot manipulation training. As data accumulates, the assistive agent gradually learns. Consequently, less human effort and attention are required, enhancing the efficiency of the data collection process. It also allows the human operator to adjust the control ratio to achieve a trade-off between manual and automated control. We conducted experiments in both simulated environments and physical real-world settings. Through user studies and quantitative evaluations, it is evident that the proposed system could enhance data collection efficiency and reduce the need for human adaptation while ensuring the collected data is of sufficient quality for downstream tasks. \textit{For more details, please refer to our webpage https://norweig1an.github.io/HAJL.github.io/.

comment: 8 pages, 6 figures

♻ ☆ AROMA: Preserving Spatial Structure for Latent PDE Modeling with Local Neural Fields

Louis Serrano, Thomas X Wang, Etienne Le Naour, Jean-Noël Vittaut, Patrick Gallinari

We present AROMA (Attentive Reduced Order Model with Attention), a framework designed to enhance the modeling of partial differential equations (PDEs) using local neural fields. Our flexible encoder-decoder architecture can obtain smooth latent representations of spatial physical fields from a variety of data types, including irregular-grid inputs and point clouds. This versatility eliminates the need for patching and allows efficient processing of diverse geometries. The sequential nature of our latent representation can be interpreted spatially and permits the use of a conditional transformer for modeling the temporal dynamics of PDEs. By employing a diffusion-based formulation, we achieve greater stability and enable longer rollouts compared to conventional MSE training. AROMA's superior performance in simulating 1D and 2D equations underscores the efficacy of our approach in capturing complex dynamical behaviors.

♻ ☆ Feature Mapping in Physics-Informed Neural Networks (PINNs)

Chengxi Zeng, Tilo Burghardt, Alberto M Gambaruto

In this paper, the training dynamics of PINNs with a feature mapping layer via the limiting Conjugate Kernel and Neural Tangent Kernel is investigated, shedding light on the convergence of PINNs; Although the commonly used Fourier-based feature mapping has achieved great success, we show its inadequacy in some physics scenarios. Via these two scopes, we propose conditionally positive definite Radial Basis Function as a better alternative. Lastly, we explore the feature mapping numerically in wide neural networks. Our empirical results reveal the efficacy of our method in diverse forward and inverse problem sets. Composing feature functions is found to be a practical way to address the expressivity and generalisability trade-off, viz., tuning the bandwidth of the kernels and the surjectivity of the feature mapping function. This simple technique can be implemented for coordinate inputs and benefits the broader PINNs research.

♻ ☆ On the Role of Context in Reading Time Prediction EMNLP 2024

Andreas Opedal, Eleanor Chodroff, Ryan Cotterell, Ethan Gotlieb Wilcox

We present a new perspective on how readers integrate context during real-time language comprehension. Our proposals build on surprisal theory, which posits that the processing effort of a linguistic unit (e.g., a word) is an affine function of its in-context information content. We first observe that surprisal is only one out of many potential ways that a contextual predictor can be derived from a language model. Another one is the pointwise mutual information (PMI) between a unit and its context, which turns out to yield the same predictive power as surprisal when controlling for unigram frequency. Moreover, both PMI and surprisal are correlated with frequency. This means that neither PMI nor surprisal contains information about context alone. In response to this, we propose a technique where we project surprisal onto the orthogonal complement of frequency, yielding a new contextual predictor that is uncorrelated with frequency. Our experiments show that the proportion of variance in reading times explained by context is a lot smaller when context is represented by the orthogonalized predictor. From an interpretability standpoint, this indicates that previous studies may have overstated the role that context has in predicting reading times.

comment: EMNLP 2024

♻ ☆ TabSeq: A Framework for Deep Learning on Tabular Data via Sequential Ordering ICPR 2024

Al Zadid Sultan Bin Habib, Kesheng Wang, Mary-Anne Hartley, Gianfranco Doretto, Donald A. Adjeroh

Effective analysis of tabular data still poses a significant problem in deep learning, mainly because features in tabular datasets are often heterogeneous and have different levels of relevance. This work introduces TabSeq, a novel framework for the sequential ordering of features, addressing the vital necessity to optimize the learning process. Features are not always equally informative, and for certain deep learning models, their random arrangement can hinder the model's learning capacity. Finding the optimum sequence order for such features could improve the deep learning models' learning process. The novel feature ordering technique we provide in this work is based on clustering and incorporates both local ordering and global ordering. It is designed to be used with a multi-head attention mechanism in a denoising autoencoder network. Our framework uses clustering to align comparable features and improve data organization. Multi-head attention focuses on essential characteristics, whereas the denoising autoencoder highlights important aspects by rebuilding from distorted inputs. This method improves the capability to learn from tabular data while lowering redundancy. Our research, demonstrating improved performance through appropriate feature sequence rearrangement using raw antibody microarray and two other real-world biomedical datasets, validates the impact of feature ordering. These results demonstrate that feature ordering can be a viable approach to improved deep learning of tabular data.

comment: This paper has been accepted for presentation at the 27th International Conference on Pattern Recognition (ICPR 2024) in Kolkata, India

♻ ☆ Pre-processing and Compression: Understanding Hidden Representation Refinement Across Imaging Domains via Intrinsic Dimension NeurIPS 2024

Nicholas Konz, Maciej A. Mazurowski

In recent years, there has been interest in how geometric properties such as intrinsic dimension (ID) of a neural network's hidden representations change through its layers, and how such properties are predictive of important model behavior such as generalization ability. However, evidence has begun to emerge that such behavior can change significantly depending on the domain of the network's training data, such as natural versus medical images. Here, we further this inquiry by exploring how the ID of a network's learned representations changes through its layers, in essence, characterizing how the network successively refines the information content of input data to be used for predictions. Analyzing eleven natural and medical image datasets across six network architectures, we find that how ID changes through the network differs noticeably between natural and medical image models. Specifically, medical image models peak in representation ID earlier in the network, implying a difference in the image features and their abstractness that are typically used for downstream tasks in these domains. Additionally, we discover a strong correlation of this peak representation ID with the ID of the data in its input space, implying that the intrinsic information content of a model's learned representations is guided by that of the data it was trained on. Overall, our findings emphasize notable discrepancies in network behavior between natural and non-natural imaging domains regarding hidden representation information content, and provide further insights into how a network's learned features are shaped by its training data.

comment: Published in NeurIPS 2024 Workshop on Scientific Methods for Understanding Deep Learning (SciForDL)

♻ ☆ Automatic AI Model Selection for Wireless Systems: Online Learning via Digital Twinning

Qiushuo Hou, Matteo Zecchin, Sangwoo Park, Yunlong Cai, Guanding Yu, Kaushik Chowdhury, Osvaldo Simeone

In modern wireless network architectures, such as O-RAN, artificial intelligence (AI)-based applications are deployed at intelligent controllers to carry out functionalities like scheduling or power control. The AI "apps" are selected on the basis of contextual information such as network conditions, topology, traffic statistics, and design goals. The mapping between context and AI model parameters is ideally done in a zero-shot fashion via an automatic model selection (AMS) mapping that leverages only contextual information without requiring any current data. This paper introduces a general methodology for the online optimization of AMS mappings. Optimizing an AMS mapping is challenging, as it requires exposure to data collected from many different contexts. Therefore, if carried out online, this initial optimization phase would be extremely time consuming. A possible solution is to leverage a digital twin of the physical system to generate synthetic data from multiple simulated contexts. However, given that the simulator at the digital twin is imperfect, a direct use of simulated data for the optimization of the AMS mapping would yield poor performance when tested in the real system. This paper proposes a novel method for the online optimization of AMS mapping that corrects for the bias of the simulator by means of limited real data collected from the physical system. Experimental results for a graph neural network-based power control app demonstrate the significant advantages of the proposed approach.

comment: submitted for a journal publication

♻ ☆ A PID-Controlled Non-Negative Tensor Factorization Model for Analyzing Missing Data in NILM

DengYu Shi

With the growing demand for energy and increased environmental awareness, Non-Intrusive Load Monitoring (NILM) has become an essential tool in smart grid and energy management. By analyzing total power load data, NILM infers the energy usage of individual appliances without the need for separate sensors, enabling real-time monitoring from a few locations. This approach helps users understand consumption patterns, enhance energy efficiency, and detect anomalies for effective energy management. However, NILM datasets often suffer from issues such as sensor failures and data loss, compromising data integrity, thereby impacting subsequent analysis and applications. Traditional imputation methods, such as linear interpolation and matrix factorization, struggle with nonlinear relationships and are sensitive to sparse data, resulting in information loss. To address these challenges, this paper proposes a Proportional-Integral-Derivative (PID) Controlled Non-Negative Latent Factorization of Tensor (PNLF) model, which dynamically adjusts parameter gradients to improve convergence, stability, and accuracy. Experimental results show that the PNLF model significantly outperforms state-of-the-art tensor completion models in both accuracy and efficiency. By addressing data loss issues, this study enhances load disaggregation precision and optimizes energy management, providing reliable data support for smart grid applications and policy formulation.

comment: 13papegs 8figures

♻ ☆ Generalized Group Data Attribution

Dan Ley, Suraj Srinivas, Shichang Zhang, Gili Rusak, Himabindu Lakkaraju

Data Attribution (DA) methods quantify the influence of individual training data points on model outputs and have broad applications such as explainability, data selection, and noisy label identification. However, existing DA methods are often computationally intensive, limiting their applicability to large-scale machine learning models. To address this challenge, we introduce the Generalized Group Data Attribution (GGDA) framework, which computationally simplifies DA by attributing to groups of training points instead of individual ones. GGDA is a general framework that subsumes existing attribution methods and can be applied to new DA techniques as they emerge. It allows users to optimize the trade-off between efficiency and fidelity based on their needs. Our empirical results demonstrate that GGDA applied to popular DA methods such as Influence Functions, TracIn, and TRAK results in upto 10x-50x speedups over standard DA methods while gracefully trading off attribution fidelity. For downstream applications such as dataset pruning and noisy label identification, we demonstrate that GGDA significantly improves computational efficiency and maintains effectiveness, enabling practical applications in large-scale machine learning scenarios that were previously infeasible.

♻ ☆ Wasserstein Gradient Flow over Variational Parameter Space for Variational Inference

Dai Hai Nguyen, Tetsuya Sakurai, Hiroshi Mamitsuka

Variational inference (VI) can be cast as an optimization problem in which the variational parameters are tuned to closely align a variational distribution with the true posterior. The optimization task can be approached through vanilla gradient descent in black-box VI or natural-gradient descent in natural-gradient VI. In this work, we reframe VI as the optimization of an objective that concerns probability distributions defined over a \textit{variational parameter space}. Subsequently, we propose Wasserstein gradient descent for tackling this optimization problem. Notably, the optimization techniques, namely black-box VI and natural-gradient VI, can be reinterpreted as specific instances of the proposed Wasserstein gradient descent. To enhance the efficiency of optimization, we develop practical methods for numerically solving the discrete gradient flows. We validate the effectiveness of the proposed methods through empirical experiments on a synthetic dataset, supplemented by theoretical analyses.

♻ ☆ Machine Unlearning in Forgettability Sequence

Junjie Chen, Qian Chen, Jian Lou, Xiaoyu Zhang, Kai Wu, Zilong Wang

Machine unlearning (MU) is becoming a promising paradigm to achieve the "right to be forgotten", where the training trace of any chosen data points could be eliminated, while maintaining the model utility on general testing samples after unlearning. With the advancement of forgetting research, many fundamental open questions remain unanswered: do different samples exhibit varying levels of difficulty in being forgotten? Further, does the sequence in which samples are forgotten, determined by their respective difficulty levels, influence the performance of forgetting algorithms? In this paper, we identify key factor affecting unlearning difficulty and the performance of unlearning algorithms. We find that samples with higher privacy risks are more likely to be unlearning, indicating that the unlearning difficulty varies among different samples which motives a more precise unlearning mode. Built upon this insight, we propose a general unlearning framework, dubbed RSU, which consists of Ranking module and SeqUnlearn module.

comment: The senior authors of the draft are not fully convinced that the novelty is significant enough for this submission compared to the latest research progress in this area. Additionally, the senior authors have identified writing issues. Based on these two reasons, we have decided to withdraw the draft from arXiv

♻ ☆ Dynamics of Moral Behavior in Heterogeneous Populations of Learning Agents AAAI

Elizaveta Tennant, Stephen Hailes, Mirco Musolesi

Growing concerns about safety and alignment of AI systems highlight the importance of embedding moral capabilities in artificial agents: a promising solution is the use of learning from experience, i.e., Reinforcement Learning. In multi-agent (social) environments, complex population-level phenomena may emerge from interactions between individual learning agents. Many of the existing studies rely on simulated social dilemma environments to study the interactions of independent learning agents; however, they tend to ignore the moral heterogeneity that is likely to be present in societies of agents in practice. For example, at different points in time a single learning agent may face opponents who are consequentialist (i.e., focused on maximizing outcomes over time), norm-based (i.e., conforming to specific norms), or virtue-based (i.e., considering a combination of different virtues). The extent to which agents' co-development may be impacted by such moral heterogeneity in populations is not well understood. In this paper, we present a study of the learning dynamics of morally heterogeneous populations interacting in a social dilemma setting. Using an Iterated Prisoner's Dilemma environment with a partner selection mechanism, we investigate the extent to which the prevalence of diverse moral agents in populations affects individual agents' learning behaviors and emergent population-level outcomes. We observe several types of non-trivial interactions between pro-social and anti-social agents, and find that certain types of moral agents are able to steer selfish agents towards more cooperative behavior.

comment: Presented at AIES 2024 (7th AAAI/ACM Conference on AI, Ethics, and Society - San Jose, CA, USA) https://ojs.aaai.org/index.php/AIES/article/view/31736

♻ ☆ Bypass Back-propagation: Optimization-based Structural Pruning for Large Language Models via Policy Gradient

Yuan Gao, Zujing Liu, Weizhong Zhang, Bo Du, Gui-Song Xia

In contrast to moderate-size neural network pruning, structural weight pruning on the Large-Language Models (LLMs) imposes a novel challenge on the efficiency of the pruning algorithms, due to the heavy computation/memory demands of the LLMs. Recent efficient LLM pruning methods typically operate at the post-training phase without the expensive weight finetuning, however, their pruning criteria often rely on heuristically hand-crafted metrics, potentially leading to suboptimal performance. We instead propose a novel optimization-based structural pruning that learns the pruning masks in a probabilistic space directly by optimizing the loss of the pruned model. To preserve the efficiency, our method eliminates the back-propagation through the LLM per se during the optimization, requiring only the forward pass of the LLM. We achieve this by learning an underlying Bernoulli distribution to sample binary pruning masks, where we decouple the Bernoulli parameters from the LLM loss, thus facilitating an efficient optimization via a policy gradient estimator without back-propagation. As a result, our method is able to 1) operate at structural granularities of channels, heads, and layers, 2) support global and heterogeneous pruning (i.e., our method automatically determines different redundancy for different layers), and 3) optionally initialize with a metric-based method (for our Bernoulli distributions). Extensive experiments on LLaMA, LLaMA-2, LLaMA-3, Vicuna, and Mistral using the C4 and WikiText2 datasets demonstrate that our method operates for 2.7 hours with around 35GB memory for the 13B models on a single A100 GPU, and our pruned models outperform the state-of-the-arts w.r.t. both perplexity and the majority of various zero-shot tasks. Codes will be released.

comment: Initially submitted on June 15, 2024, this version mainly changed the title, and added several experiments: such as 1) experiments on LLaMA-3, Mistral, 2) additional baseline methods (i.e., Bosai -- Everybody Prune Now), and 3) post-pruning finetuned performance (i.e., first prune then finetune)

♻ ☆ Toward Routing River Water in Land Surface Models with Recurrent Neural Networks

Mauricio Lima, Katherine Deck, Oliver R. A. Dunbar, Tapio Schneider

Machine learning is playing an increasing role in hydrology, supplementing or replacing physics-based models. One notable example is the use of recurrent neural networks (RNNs) for forecasting streamflow given observed precipitation and geographic characteristics. Training of such a model over the continental United States (CONUS) demonstrated that a single set of model parameters can be used across independent catchments, and that RNNs can outperform physics-based models. In this work, we take a next step and study the performance of RNNs for river routing in land surface models (LSMs). Instead of observed precipitation, the LSM-RNN uses instantaneous runoff calculated from physics-based models as an input. We train the model with data from river basins spanning the globe and test using historical streamflow measurements. The model demonstrates skill at generalization across basins (predicting streamflow in catchments not used in training) and across time (predicting streamflow during years not used in training). We compare the predictions from the LSM-RNN to an existing physics-based model calibrated with a similar dataset and find that the LSM-RNN outperforms the physics-based model. Our results show that RNNs are effective for global streamflow prediction from runoff inputs and motivate the development of complete routing models that can capture nested sub-basis connections.

comment: 32 pages, 11 figures; submitted in HESS (EGU) with CCBY license

♻ ☆ Is the MMI Criterion Necessary for Interpretability? Degenerating Non-causal Features to Plain Noise for Self-Rationalization NeurIPS 2024

Wei Liu, Zhiying Deng, Zhongyu Niu, Jun Wang, Haozhao Wang, YuanKai Zhang, Ruixuan Li

An important line of research in the field of explainability is to extract a small subset of crucial rationales from the full input. The most widely used criterion for rationale extraction is the maximum mutual information (MMI) criterion. However, in certain datasets, there are spurious features non-causally correlated with the label and also get high mutual information, complicating the loss landscape of MMI. Although some penalty-based methods have been developed to penalize the spurious features (e.g., invariance penalty, intervention penalty, etc) to help MMI work better, these are merely remedial measures. In the optimization objectives of these methods, spurious features are still distinguished from plain noise, which hinders the discovery of causal rationales. This paper aims to develop a new criterion that treats spurious features as plain noise, allowing the model to work on datasets rich in spurious features as if it were working on clean datasets, thereby making rationale extraction easier. We theoretically observe that removing either plain noise or spurious features from the input does not alter the conditional distribution of the remaining components relative to the task label. However, significant changes in the conditional distribution occur only when causal features are eliminated. Based on this discovery, the paper proposes a criterion for \textbf{M}aximizing the \textbf{R}emaining \textbf{D}iscrepancy (MRD). Experiments on six widely used datasets show that our MRD criterion improves rationale quality (measured by the overlap with human-annotated rationales) by up to $10.4\%$ as compared to several recent competitive MMI variants. Code: \url{https://github.com/jugechengzi/Rationalization-MRD}.

comment: Accepted at NeurIPS 2024. arXiv admin note: text overlap with arXiv:2309.13391

♻ ☆ Selection-p: Self-Supervised Task-Agnostic Prompt Compression for Faithfulness and Transferability EMNLP 2024

Tsz Ting Chung, Leyang Cui, Lemao Liu, Xinting Huang, Shuming Shi, Dit-Yan Yeung

Large Language Models (LLMs) have demonstrated impressive capabilities in a wide range of natural language processing tasks when leveraging in-context learning. To mitigate the additional computational and financial costs associated with in-context learning, several prompt compression methods have been proposed to compress the in-context learning prompts. Despite their success, these methods face challenges with transferability due to model-specific compression, or rely on external training data, such as GPT-4. In this paper, we investigate the ability of LLMs to develop a unified compression method that discretizes uninformative tokens, utilizing a self-supervised pre-training technique. By introducing a small number of parameters during the continual pre-training, the proposed Selection-p produces a probability for each input token, indicating whether to preserve or discard it. Experiments show Selection-p achieves state-of-the-art performance across numerous classification tasks, achieving compression rates of up to 10 times while experiencing only a marginal 0.8% decrease in performance. Moreover, it exhibits superior transferability to different models compared to prior work. Additionally, we further analyze how Selection-p helps maintain performance on in-context learning with long contexts.

comment: 14 pages, 5 figures, 10 tables, EMNLP 2024 Findings

♻ ☆ Neural Collapse versus Low-rank Bias: Is Deep Neural Collapse Really Optimal?

Peter Súkeník, Marco Mondelli, Christoph Lampert

Deep neural networks (DNNs) exhibit a surprising structure in their final layer known as neural collapse (NC), and a growing body of works has currently investigated the propagation of neural collapse to earlier layers of DNNs -- a phenomenon called deep neural collapse (DNC). However, existing theoretical results are restricted to special cases: linear models, only two layers or binary classification. In contrast, we focus on non-linear models of arbitrary depth in multi-class classification and reveal a surprising qualitative shift. As soon as we go beyond two layers or two classes, DNC stops being optimal for the deep unconstrained features model (DUFM) -- the standard theoretical framework for the analysis of collapse. The main culprit is a low-rank bias of multi-layer regularization schemes: this bias leads to optimal solutions of even lower rank than the neural collapse. We support our theoretical findings with experiments on both DUFM and real data, which show the emergence of the low-rank structure in the solution found by gradient descent.

♻ ☆ Towards Next-Level Post-Training Quantization of Hyper-Scale Transformers NeurIPS 2024

Junhan Kim, Chungman Lee, Eulrang Cho, Kyungphil Park, Ho-young Kim, Joonyoung Kim, Yongkweon Jeon

With the increasing complexity of generative AI models, post-training quantization (PTQ) has emerged as a promising solution for deploying hyper-scale models on edge devices such as mobile and TVs. Existing PTQ schemes, however, consume considerable time and resources, which could be a bottleneck in real situations where frequent model updates and multiple hyperparameter tunings are required. As a cost-effective alternative, learning-free PTQ schemes have been proposed. However, the performance is somewhat limited because they cannot consider the inter-layer dependency within the attention module, which is a significant feature of Transformers. In this paper, we thus propose a novel PTQ algorithm that balances accuracy and efficiency. The key idea of the proposed algorithm called aespa is to perform quantization layer-wise for efficiency while targeting attention-wise reconstruction to consider the cross-layer dependency. Through extensive experiments on various language models and complexity analysis, we demonstrate that aespa is accurate and efficient in quantizing Transformer models.

comment: Accepted to NeurIPS 2024

♻ ☆ ProbTS: Benchmarking Point and Distributional Forecasting across Diverse Prediction Horizons NeurIPS 2024

Jiawen Zhang, Xumeng Wen, Zhenwei Zhang, Shun Zheng, Jia Li, Jiang Bian

Delivering precise point and distributional forecasts across a spectrum of prediction horizons represents a significant and enduring challenge in the application of time-series forecasting within various industries. Prior research on developing deep learning models for time-series forecasting has often concentrated on isolated aspects, such as long-term point forecasting or short-term probabilistic estimations. This narrow focus may result in skewed methodological choices and hinder the adaptability of these models to uncharted scenarios. While there is a rising trend in developing universal forecasting models, a thorough understanding of their advantages and drawbacks, especially regarding essential forecasting needs like point and distributional forecasts across short and long horizons, is still lacking. In this paper, we present ProbTS, a benchmark tool designed as a unified platform to evaluate these fundamental forecasting needs and to conduct a rigorous comparative analysis of numerous cutting-edge studies from recent years. We dissect the distinctive data characteristics arising from disparate forecasting requirements and elucidate how these characteristics can skew methodological preferences in typical research trajectories, which often fail to fully accommodate essential forecasting needs. Building on this, we examine the latest models for universal time-series forecasting and discover that our analyses of methodological strengths and weaknesses are also applicable to these universal models. Finally, we outline the limitations inherent in current research and underscore several avenues for future exploration.

comment: NeurIPS 2024 Datasets and Benchmarks Track

♻ ☆ Few-sample Variational Inference of Bayesian Neural Networks with Arbitrary Nonlinearities

David J. Schodt

Bayesian Neural Networks (BNNs) extend traditional neural networks to provide uncertainties associated with their outputs. On the forward pass through a BNN, predictions (and their uncertainties) are made either by Monte Carlo sampling network weights from the learned posterior or by analytically propagating statistical moments through the network. Though flexible, Monte Carlo sampling is computationally expensive and can be infeasible or impractical under resource constraints or for large networks. While moment propagation can ameliorate the computational costs of BNN inference, it can be difficult or impossible for networks with arbitrary nonlinearities, thereby restricting the possible set of network layers permitted with such a scheme. In this work, we demonstrate a simple yet effective approach for propagating statistical moments through arbitrary nonlinearities with only 3 deterministic samples, enabling few-sample variational inference of BNNs without restricting the set of network layers used. Furthermore, we leverage this approach to demonstrate a novel nonlinear activation function that we use to inject physics-informed prior information into output nodes of a BNN.

comment: Comment 1: Fixed plot markers in figure 6 to match legend and to improve grayscale appearance Comment 2: Fixed mistyped value for optimizer learning rate

♻ ☆ Generating Less Certain Adversarial Examples Improves Robust Generalization

Minxing Zhang, Michael Backes, Xiao Zhang

This paper revisits the robust overfitting phenomenon of adversarial training. Observing that models with better robust generalization performance are less certain in predicting adversarially generated training inputs, we argue that overconfidence in predicting adversarial examples is a potential cause. Therefore, we hypothesize that generating less certain adversarial examples improves robust generalization, and propose a formal definition of adversarial certainty that captures the variance of the model's predicted logits on adversarial examples. Our theoretical analysis of synthetic distributions characterizes the connection between adversarial certainty and robust generalization. Accordingly, built upon the notion of adversarial certainty, we develop a general method to search for models that can generate training-time adversarial inputs with reduced certainty, while maintaining the model's capability in distinguishing adversarial examples. Extensive experiments on image benchmarks demonstrate that our method effectively learns models with consistently improved robustness and mitigates robust overfitting, confirming the importance of generating less certain adversarial examples for robust generalization.

comment: Published in Transactions of Machine Learning Research (TMLR)

♻ ☆ The Ultimate Guide to Fine-Tuning LLMs from Basics to Breakthroughs: An Exhaustive Review of Technologies, Research, Best Practices, Applied Research Challenges and Opportunities

Venkatesh Balavadhani Parthasarathy, Ahtsham Zafar, Aafaq Khan, Arsalan Shahid

This report examines the fine-tuning of Large Language Models (LLMs), integrating theoretical insights with practical applications. It outlines the historical evolution of LLMs from traditional Natural Language Processing (NLP) models to their pivotal role in AI. A comparison of fine-tuning methodologies, including supervised, unsupervised, and instruction-based approaches, highlights their applicability to different tasks. The report introduces a structured seven-stage pipeline for fine-tuning LLMs, spanning data preparation, model initialization, hyperparameter tuning, and model deployment. Emphasis is placed on managing imbalanced datasets and optimization techniques. Parameter-efficient methods like Low-Rank Adaptation (LoRA) and Half Fine-Tuning are explored for balancing computational efficiency with performance. Advanced techniques such as memory fine-tuning, Mixture of Experts (MoE), and Mixture of Agents (MoA) are discussed for leveraging specialized networks and multi-agent collaboration. The report also examines novel approaches like Proximal Policy Optimization (PPO) and Direct Preference Optimization (DPO), which align LLMs with human preferences, alongside pruning and routing optimizations to improve efficiency. Further sections cover validation frameworks, post-deployment monitoring, and inference optimization, with attention to deploying LLMs on distributed and cloud-based platforms. Emerging areas such as multimodal LLMs, fine-tuning for audio and speech, and challenges related to scalability, privacy, and accountability are also addressed. This report offers actionable insights for researchers and practitioners navigating LLM fine-tuning in an evolving landscape.

♻ ☆ An Electrocardiogram Foundation Model Built on over 10 Million Recordings with External Evaluation across Multiple Domains

Jun Li, Aaron Aguirre, Junior Moura, Che Liu, Lanhai Zhong, Chenxi Sun, Gari Clifford, Brandon Westover, Shenda Hong

Artificial intelligence (AI) has demonstrated significant potential in ECG analysis and cardiovascular disease assessment. Recently, foundation models have played a remarkable role in advancing medical AI. The development of an ECG foundation model holds the promise of elevating AI-ECG research to new heights. However, building such a model faces several challenges, including insufficient database sample sizes and inadequate generalization across multiple domains. Additionally, there is a notable performance gap between single-lead and multi-lead ECG analyses. We introduced an ECG Foundation Model (ECGFounder), a general-purpose model that leverages real-world ECG annotations from cardiology experts to broaden the diagnostic capabilities of ECG analysis. ECGFounder was trained on over 10 million ECGs with 150 label categories from the Harvard-Emory ECG Database, enabling comprehensive cardiovascular disease diagnosis through ECG analysis. The model is designed to be both an effective out-of-the-box solution, and a to be fine-tunable for downstream tasks, maximizing usability. Importantly, we extended its application to lower rank ECGs, and arbitrary single-lead ECGs in particular. ECGFounder is applicable to supporting various downstream tasks in mobile monitoring scenarios. Experimental results demonstrate that ECGFounder achieves expert-level performance on internal validation sets, with AUROC exceeding 0.95 for eighty diagnoses. It also shows strong classification performance and generalization across various diagnoses on external validation sets. When fine-tuned, ECGFounder outperforms baseline models in demographic analysis, clinical event detection, and cross-modality cardiac rhythm diagnosis. The trained model and data will be publicly released upon publication through the bdsp.io. Our code is available at https://github.com/bdsp-core/ECGFounder

comment: working in progress

♻ ★ Any2Point: Empowering Any-modality Large Models for Efficient 3D Understanding

Yiwen Tang, Ray Zhang, Jiaming Liu, Zoey Guo, Dong Wang, Zhigang Wang, Bin Zhao, Shanghang Zhang, Peng Gao, Hongsheng Li, Xuelong Li

Large foundation models have recently emerged as a prominent focus of interest, attaining superior performance in widespread scenarios. Due to the scarcity of 3D data, many efforts have been made to adapt pre-trained transformers from vision to 3D domains. However, such 2D-to-3D approaches are still limited, due to the potential loss of spatial geometries and high computation cost. More importantly, their frameworks are mainly designed for 2D models, lacking a general any-to-3D paradigm. In this paper, we introduce Any2Point, a parameter-efficient method to empower any-modality large models (vision, language, audio) for 3D understanding. Given a frozen transformer from any source modality, we propose a 3D-to-any (1D or 2D) virtual projection strategy that correlates the input 3D points to the original 1D or 2D positions within the source modality. This mechanism enables us to assign each 3D token with a positional encoding paired with the pre-trained model, which avoids 3D geometry loss caused by the true projection and better motivates the transformer for 3D learning with 1D/2D positional priors. Then, within each transformer block, we insert an any-to-3D guided adapter module for parameter-efficient fine-tuning. The adapter incorporates prior spatial knowledge from the source modality to guide the local feature aggregation of 3D tokens, compelling the semantic adaption of any-modality transformers. We conduct extensive experiments to showcase the effectiveness and efficiency of our method. Code and models are released at https://github.com/Ivan-Tang-3D/Any2Point.

comment: Code and models are released at https://github.com/Ivan-Tang-3D/Any2Point

♻ ☆ Point-PEFT: Parameter-Efficient Fine-Tuning for 3D Pre-trained Models

Yiwen Tang, Ray Zhang, Zoey Guo, Dong Wang, Zhigang Wang, Bin Zhao, Xuelong Li

The popularity of pre-trained large models has revolutionized downstream tasks across diverse fields, such as language, vision, and multi-modality. To minimize the adaption cost for downstream tasks, many Parameter-Efficient Fine-Tuning (PEFT) techniques are proposed for language and 2D image pre-trained models. However, the specialized PEFT method for 3D pre-trained models is still under-explored. To this end, we introduce Point-PEFT, a novel framework for adapting point cloud pre-trained models with minimal learnable parameters. Specifically, for a pre-trained 3D model, we freeze most of its parameters, and only tune the newly added PEFT modules on downstream tasks, which consist of a Point-prior Prompt and a Geometry-aware Adapter. The Point-prior Prompt adopts a set of learnable prompt tokens, for which we propose to construct a memory bank with domain-specific knowledge, and utilize a parameter-free attention to enhance the prompt tokens. The Geometry-aware Adapter aims to aggregate point cloud features within spatial neighborhoods to capture fine-grained geometric information through local interactions. Extensive experiments indicate that our Point-PEFT can achieve better performance than the full fine-tuning on various downstream tasks, while using only 5% of the trainable parameters, demonstrating the efficiency and effectiveness of our approach. Code is released at https://github.com/Ivan-Tang-3D/Point-PEFT.

comment: The specialized PEFT framework for 3D pre-trained models, which achieves competitive performance to full fine-tuning, and significantly reduces the computational resources. Project page: https://github.com/Ivan-Tang-3D/Point-PEFT

♻ ☆ A physics-informed neural network framework for modeling obstacle-related equations

Hamid El Bahja, Jan Christian Hauffen, Peter Jung, Bubacarr Bah, Issa Karambal

Deep learning has been highly successful in some applications. Nevertheless, its use for solving partial differential equations (PDEs) has only been of recent interest with current state-of-the-art machine learning libraries, e.g., TensorFlow or PyTorch. Physics-informed neural networks (PINNs) are an attractive tool for solving partial differential equations based on sparse and noisy data. Here extend PINNs to solve obstacle-related PDEs which present a great computational challenge because they necessitate numerical methods that can yield an accurate approximation of the solution that lies above a given obstacle. The performance of the proposed PINNs is demonstrated in multiple scenarios for linear and nonlinear PDEs subject to regular and irregular obstacles.

♻ ☆ Flat Posterior Does Matter For Bayesian Model Averaging

Sungjun Lim, Jeyoon Yeom, Sooyon Kim, Hoyoon Byun, Jinho Kang, Yohan Jung, Jiyoung Jung, Kyungwoo Song

Bayesian neural network (BNN) approximates the posterior distribution of model parameters and utilizes the posterior for prediction via Bayesian Model Averaging (BMA). The quality of the posterior approximation is critical for achieving accurate and robust predictions. It is known that flatness in the loss landscape is strongly associated with generalization performance, and it necessitates consideration to improve the quality of the posterior approximation. In this work, we empirically demonstrate that BNNs often struggle to capture the flatness. Moreover, we provide both experimental and theoretical evidence showing that BMA can be ineffective without ensuring flatness. To address this, we propose Sharpness-Aware Bayesian Model Averaging (SA-BMA), a novel optimizer that seeks flat posteriors by calculating divergence in the parameter space. SA-BMA aligns with the intrinsic nature of BNN and the generalized version of existing sharpness-aware optimizers for DNN. In addition, we suggest a Bayesian Transfer Learning scheme to efficiently leverage pre-trained DNN. We validate the efficacy of SA-BMA in enhancing generalization performance in few-shot classification and distribution shift by ensuring flat posterior.

♻ ☆ Amortized Planning with Large-Scale Transformers: A Case Study on Chess

Anian Ruoss, Grégoire Delétang, Sourabh Medapati, Jordi Grau-Moya, Li Kevin Wenliang, Elliot Catt, John Reid, Cannada A. Lewis, Joel Veness, Tim Genewein

This paper uses chess, a landmark planning problem in AI, to assess transformers' performance on a planning task where memorization is futile $\unicode{x2013}$ even at a large scale. To this end, we release ChessBench, a large-scale benchmark dataset of 10 million chess games with legal move and value annotations (15 billion data points) provided by Stockfish 16, the state-of-the-art chess engine. We train transformers with up to 270 million parameters on ChessBench via supervised learning and perform extensive ablations to assess the impact of dataset size, model size, architecture type, and different prediction targets (state-values, action-values, and behavioral cloning). Our largest models learn to predict action-values for novel boards quite accurately, implying highly non-trivial generalization. Despite performing no explicit search, our resulting chess policy solves challenging chess puzzles and achieves a surprisingly strong Lichess blitz Elo of 2895 against humans (grandmaster level). We also compare to Leela Chess Zero and AlphaZero (trained without supervision via self-play) with and without search. We show that, although a remarkably good approximation of Stockfish's search-based algorithm can be distilled into large-scale transformers via supervised learning, perfect distillation is still beyond reach, thus making ChessBench well-suited for future research.

♻ ☆ NETS: A Non-Equilibrium Transport Sampler

Michael S. Albergo, Eric Vanden-Eijnden

We propose an algorithm, termed the Non-Equilibrium Transport Sampler (NETS), to sample from unnormalized probability distributions. NETS can be viewed as a variant of annealed importance sampling (AIS) based on Jarzynski's equality, in which the stochastic differential equation used to perform the non-equilibrium sampling is augmented with an additional learned drift term that lowers the impact of the unbiasing weights used in AIS. We show that this drift is the minimizer of a variety of objective functions, which can all be estimated in an unbiased fashion without backpropagating through solutions of the stochastic differential equations governing the sampling. We also prove that some these objectives control the Kullback-Leibler divergence of the estimated distribution from its target. NETS is shown to be unbiased and, in addition, has a tunable diffusion coefficient which can be adjusted post-training to maximize the effective sample size. We demonstrate the efficacy of the method on standard benchmarks, high-dimensional Gaussian mixture distributions, and a model from statistical lattice field theory, for which it surpasses the performances of related work and existing baselines.

♻ ☆ Deep Multimodal Learning with Missing Modality: A Survey

Renjie Wu, Hu Wang, Hsiang-Ting Chen, Gustavo Carneiro

During multimodal model training and testing, certain data modalities may be absent due to sensor limitations, cost constraints, privacy concerns, or data loss, negatively affecting performance. Multimodal learning techniques designed to handle missing modalities can mitigate this by ensuring model robustness even when some modalities are unavailable. This survey reviews recent progress in Multimodal Learning with Missing Modality (MLMM), focusing on deep learning methods. It provides the first comprehensive survey that covers the motivation and distinctions between MLMM and standard multimodal learning setups, followed by a detailed analysis of current methods, applications, and datasets, concluding with challenges and future directions.

comment: Submitted to ACM Computing Surveys

♻ ☆ The Art of Imitation: Learning Long-Horizon Manipulation Tasks from Few Demonstrations

Jan Ole von Hartz, Tim Welschehold, Abhinav Valada, Joschka Boedecker

Task Parametrized Gaussian Mixture Models (TP-GMM) are a sample-efficient method for learning object-centric robot manipulation tasks. However, there are several open challenges to applying TP-GMMs in the wild. In this work, we tackle three crucial challenges synergistically. First, end-effector velocities are non-Euclidean and thus hard to model using standard GMMs. We thus propose to factorize the robot's end-effector velocity into its direction and magnitude, and model them using Riemannian GMMs. Second, we leverage the factorized velocities to segment and sequence skills from complex demonstration trajectories. Through the segmentation, we further align skill trajectories and hence leverage time as a powerful inductive bias. Third, we present a method to automatically detect relevant task parameters per skill from visual observations. Our approach enables learning complex manipulation tasks from just five demonstrations while using only RGB-D observations. Extensive experimental evaluations on RLBench demonstrate that our approach achieves state-of-the-art performance with 20-fold improved sample efficiency. Our policies generalize across different environments, object instances, and object positions, while the learned skills are reusable.

♻ ☆ Shotluck Holmes: A Family of Efficient Small-Scale Large Language Vision Models For Video Captioning and Summarization

Richard Luo, Austin Peng, Adithya Vasudev, Rishabh Jain

Video is an increasingly prominent and information-dense medium, yet it poses substantial challenges for language models. A typical video consists of a sequence of shorter segments, or shots, that collectively form a coherent narrative. Each shot is analogous to a word in a sentence where multiple data streams of information (such as visual and auditory data) must be processed simultaneously. Comprehension of the entire video requires not only understanding the visual-audio information of each shot but also requires that the model links the ideas between each shot to generate a larger, all-encompassing story. Despite significant progress in the field, current works often overlook videos' more granular shot-by-shot semantic information. In this project, we propose a family of efficient large language vision models (LLVMs) to boost video summarization and captioning called Shotluck Holmes. By leveraging better pretraining and data collection strategies, we extend the abilities of existing small LLVMs from being able to understand a picture to being able to understand a sequence of frames. Specifically, we show that Shotluck Holmes achieves better performance than state-of-the-art results on the Shot2Story video captioning and summary task with significantly smaller and more computationally efficient models.

♻ ☆ HYGENE: A Diffusion-based Hypergraph Generation Method

Dorian Gailhard, Enzo Tartaglione, Lirida Naviner, Jhony H. Giraldo

Hypergraphs are powerful mathematical structures that can model complex, high-order relationships in various domains, including social networks, bioinformatics, and recommender systems. However, generating realistic and diverse hypergraphs remains challenging due to their inherent complexity and lack of effective generative models. In this paper, we introduce a diffusion-based Hypergraph Generation (HYGENE) method that addresses these challenges through a progressive local expansion approach. HYGENE works on the bipartite representation of hypergraphs, starting with a single pair of connected nodes and iteratively expanding it to form the target hypergraph. At each step, nodes and hyperedges are added in a localized manner using a denoising diffusion process, which allows for the construction of the global structure before refining local details. Our experiments demonstrated the effectiveness of HYGENE, proving its ability to closely mimic a variety of properties in hypergraphs. To the best of our knowledge, this is the first attempt to employ deep learning models for hypergraph generation, and our work aims to lay the groundwork for future research in this area.

comment: arXiv admin note: text overlap with arXiv:2312.11529 by other authors

♻ ☆ Exact full-RSB SAT/UNSAT transition in infinitely wide two-layer neural networks

Brandon L. Annesi, Enrico M. Malatesta, Francesco Zamponi

We analyze the problem of storing random pattern-label associations using two classes of continuous non-convex weights models, namely the perceptron with negative margin and an infinite-width two-layer neural network with non-overlapping receptive fields and generic activation function. Using a full-RSB ansatz we compute the exact value of the SAT/UNSAT transition. Furthermore, in the case of the negative perceptron we show that the overlap distribution of typical states displays an overlap gap (a disconnected support) in certain regions of the phase diagram defined by the value of the margin and the density of patterns to be stored. This implies that some recent theorems that ensure convergence of Approximate Message Passing (AMP) based algorithms to capacity are not applicable. Finally, we show that Gradient Descent is not able to reach the maximal capacity, irrespectively of the presence of an overlap gap for typical states. This finding, similarly to what occurs in binary weight models, suggests that gradient-based algorithms are biased towards highly atypical states, whose inaccessibility determines the algorithmic threshold.

comment: 38 pages, 12 figures

♻ ☆ Adaptive Contrastive Search: Uncertainty-Guided Decoding for Open-Ended Text Generation EMNLP 2024

Esteban Garces Arias, Julian Rodemann, Meimingwei Li, Christian Heumann, Matthias Aßenmacher

Decoding from the output distributions of large language models to produce high-quality text is a complex challenge in language modeling. Various approaches, such as beam search, sampling with temperature, $k-$sampling, nucleus $p-$sampling, typical decoding, contrastive decoding, and contrastive search, have been proposed to address this problem, aiming to improve coherence, diversity, as well as resemblance to human-generated text. In this study, we introduce adaptive contrastive search, a novel decoding strategy extending contrastive search by incorporating an adaptive degeneration penalty, guided by the estimated uncertainty of the model at each generation step. This strategy is designed to enhance both the creativity and diversity of the language modeling process while at the same time producing coherent and high-quality generated text output. Our findings indicate performance enhancement in both aspects, across different model architectures and datasets, underscoring the effectiveness of our method in text generation tasks. Our code base, datasets, and models are publicly available.

comment: Accepted to EMNLP 2024 Findings

♻ ☆ Provable Acceleration of Nesterov's Accelerated Gradient for Rectangular Matrix Factorization and Linear Neural Networks

Zhenghao Xu, Yuqing Wang, Tuo Zhao, Rachel Ward, Molei Tao

We study the convergence rate of first-order methods for rectangular matrix factorization, which is a canonical nonconvex optimization problem. Specifically, given a rank-$r$ matrix $\mathbf{A}\in\mathbb{R}^{m\times n}$, we prove that gradient descent (GD) can find a pair of $\epsilon$-optimal solutions $\mathbf{X}_T\in\mathbb{R}^{m\times d}$ and $\mathbf{Y}_T\in\mathbb{R}^{n\times d}$, where $d\geq r$, satisfying $\lVert\mathbf{X}_T\mathbf{Y}_T^\top-\mathbf{A}\rVert_\mathrm{F}\leq\epsilon\lVert\mathbf{A}\rVert_\mathrm{F}$ in $T=O(\kappa^2\log\frac{1}{\epsilon})$ iterations with high probability, where $\kappa$ denotes the condition number of $\mathbf{A}$. Furthermore, we prove that Nesterov's accelerated gradient (NAG) attains an iteration complexity of $O(\kappa\log\frac{1}{\epsilon})$, which is the best-known bound of first-order methods for rectangular matrix factorization. Different from small balanced random initialization in the existing literature, we adopt an unbalanced initialization, where $\mathbf{X}_0$ is large and $\mathbf{Y}_0$ is $0$. Moreover, our initialization and analysis can be further extended to linear neural networks, where we prove that NAG can also attain an accelerated linear convergence rate. In particular, we only require the width of the network to be greater than or equal to the rank of the output label matrix. In contrast, previous results achieving the same rate require excessive widths that additionally depend on the condition number and the rank of the input data matrix.

comment: 30 pages (checklist included), fix typos

♻ ☆ A Kernelizable Primal-Dual Formulation of the Multilinear Singular Value Decomposition

Frederiek Wesel, Kim Batselier

The ability to express a learning task in terms of a primal and a dual optimization problem lies at the core of a plethora of machine learning methods. For example, Support Vector Machine (SVM), Least-Squares Support Vector Machine (LS-SVM), Ridge Regression (RR), Lasso Regression (LR), Principal Component Analysis (PCA), and more recently Singular Value Decomposition (SVD) have all been defined either in terms of primal weights or in terms of dual Lagrange multipliers. The primal formulation is computationally advantageous in the case of large sample size while the dual is preferred for high-dimensional data. Crucially, said learning problems can be made nonlinear through the introduction of a feature map in the primal problem, which corresponds to applying the kernel trick in the dual. In this paper we derive a primal-dual formulation of the Multilinear Singular Value Decomposition (MLSVD), which recovers as special cases both PCA and SVD. Besides enabling computational gains through the derived primal formulation, we propose a nonlinear extension of the MLSVD using feature maps, which results in a dual problem where a kernel tensor arises. We discuss potential applications in the context of signal analysis and deep learning.

♻ ☆ Why Transformers Need Adam: A Hessian Perspective

Yushun Zhang, Congliang Chen, Tian Ding, Ziniu Li, Ruoyu Sun, Zhi-Quan Luo

SGD performs worse than Adam by a significant margin on Transformers, but the reason remains unclear. In this work, we provide an explanation through the lens of Hessian: (i) Transformers are "heterogeneous": the Hessian spectrum across parameter blocks vary dramatically, a phenomenon we call "block heterogeneity"; (ii) Heterogeneity hampers SGD: SGD performs worse than Adam on problems with block heterogeneity. To validate (i) and (ii), we check various Transformers, CNNs, MLPs, and quadratic problems, and find that SGD can perform on par with Adam on problems without block heterogeneity, but performs worse than Adam when the heterogeneity exists. Our initial theoretical analysis indicates that SGD performs worse because it applies one single learning rate to all blocks, which cannot handle the heterogeneity among blocks. This limitation could be ameliorated if we use coordinate-wise learning rates, as designed in Adam.

comment: Advances in Neural Information Processing Systems, 2024

♻ ☆ On Uncertainty Quantification for Near-Bayes Optimal Algorithms

Ziyu Wang, Chris Holmes

Bayesian modelling allows for the quantification of predictive uncertainty which is crucial in safety-critical applications. Yet for many machine learning (ML) algorithms, it is difficult to construct or implement their Bayesian counterpart. In this work we present a promising approach to address this challenge, based on the hypothesis that commonly used ML algorithms are efficient across a wide variety of tasks and may thus be near Bayes-optimal w.r.t. an unknown task distribution. We prove that it is possible to recover the Bayesian posterior defined by the task distribution, which is unknown but optimal in this setting, by building a martingale posterior using the algorithm. We further propose a practical uncertainty quantification method that apply to general ML algorithms. Experiments based on a variety of non-NN and NN algorithms demonstrate the efficacy of our method.

♻ ☆ CaTs and DAGs: Integrating Directed Acyclic Graphs with Transformers and Fully-Connected Neural Networks for Causally Constrained Predictions

Matthew J. Vowels, Mathieu Rochat, Sina Akbari

Artificial Neural Networks (ANNs), including fully-connected networks and transformers, are highly flexible and powerful function approximators, widely applied in fields like computer vision and natural language processing. However, their inability to inherently respect causal structures can limit their robustness, making them vulnerable to covariate shift and difficult to interpret/explain. This poses significant challenges for their reliability in real-world applications. In this paper, we introduce Causal Fully-Connected Neural Networks (CFCNs) and Causal Transformers (CaTs), two general model families designed to operate under predefined causal constraints, as specified by a Directed Acyclic Graph (DAG). These models retain the powerful function approximation abilities of traditional neural networks while adhering to the underlying structural constraints, improving robustness, reliability, and interpretability at inference time. This approach opens new avenues for deploying neural networks in more demanding, real-world scenarios where robustness and explainability is critical.

♻ ☆ QUIS: Question-guided Insights Generation for Automated Exploratory Data Analysis EMNLP 2024

Abhijit Manatkar, Ashlesha Akella, Parthivi Gupta, Krishnasuri Narayanam

Discovering meaningful insights from a large dataset, known as Exploratory Data Analysis (EDA), is a challenging task that requires thorough exploration and analysis of the data. Automated Data Exploration (ADE) systems use goal-oriented methods with Large Language Models and Reinforcement Learning towards full automation. However, these methods require human involvement to anticipate goals that may limit insight extraction, while fully automated systems demand significant computational resources and retraining for new datasets. We introduce QUIS, a fully automated EDA system that operates in two stages: insight generation (ISGen) driven by question generation (QUGen). The QUGen module generates questions in iterations, refining them from previous iterations to enhance coverage without human intervention or manually curated examples. The ISGen module analyzes data to produce multiple relevant insights in response to each question, requiring no prior training and enabling QUIS to adapt to new datasets.

comment: Accepted for EMNLP 2024 Industry Track

♻ ☆ A Human-in-the-Loop Fairness-Aware Model Selection Framework for Complex Fairness Objective Landscapes

Jake Robertson, Thorsten Schmidt, Frank Hutter, Noor Awad

Fairness-aware Machine Learning (FairML) applications are often characterized by complex social objectives and legal requirements, frequently involving multiple, potentially conflicting notions of fairness. Despite the well-known Impossibility Theorem of Fairness and extensive theoretical research on the statistical and socio-technical trade-offs between fairness metrics, many FairML tools still optimize or constrain for a single fairness objective. However, this one-sided optimization can inadvertently lead to violations of other relevant notions of fairness. In this socio-technical and empirical study, we frame fairness as a many-objective (MaO) problem by treating fairness metrics as conflicting objectives. We introduce ManyFairHPO, a human-in-the-loop, fairness-aware model selection framework that enables practitioners to effectively navigate complex and nuanced fairness objective landscapes. ManyFairHPO aids in the identification, evaluation, and balancing of fairness metric conflicts and their related social consequences, leading to more informed and socially responsible model-selection decisions. Through a comprehensive empirical evaluation and a case study on the Law School Admissions problem, we demonstrate the effectiveness of ManyFairHPO in balancing multiple fairness objectives, mitigating risks such as self-fulfilling prophecies, and providing interpretable insights to guide stakeholders in making fairness-aware modeling decisions.

♻ ☆ Lightweight Correlation-Aware Table Compression NeurIPS 2024

Mihail Stoian, Alexander van Renen, Jan Kobiolka, Ping-Lin Kuo, Josif Grabocka, Andreas Kipf

The growing adoption of data lakes for managing relational data necessitates efficient, open storage formats that provide high scan performance and competitive compression ratios. While existing formats achieve fast scans through lightweight encoding techniques, they have reached a plateau in terms of minimizing storage footprint. Recently, correlation-aware compression schemes have been shown to reduce file sizes further. Yet, current approaches either incur significant scan overheads or require manual specification of correlations, limiting their practicability. We present $\texttt{Virtual}$, a framework that integrates seamlessly with existing open formats to automatically leverage data correlations, achieving substantial compression gains while having minimal scan performance overhead. Experiments on data-gov datasets show that $\texttt{Virtual}$ reduces file sizes by up to 40% compared to Apache Parquet.

comment: Third Table Representation Learning Workshop (TRL @ NeurIPS 2024)

♻ ☆ A Unified Approach to Extract Interpretable Rules from Tree Ensembles via Integer Programming

Lorenzo Bonasera, Emilio Carrizosa

Tree ensemble methods represent a popular machine learning model, known for their effectiveness in supervised classification and regression tasks. Their performance derives from aggregating predictions of multiple decision trees, which are renowned for their interpretability properties. However, tree ensemble methods do not reliably exhibit interpretable output. Our work aims to extract an optimized list of rules from a trained tree ensemble, providing the user with a condensed, interpretable model that retains most of the predictive power of the full model. Our approach consists of solving a clean and neat set partitioning problem formulated through Integer Programming. The proposed method works with either tabular or time series data, for both classification and regression tasks, and does not require parameter tuning under the most common setting. Through rigorous computational experiments, we offer statistically significant evidence that our method is competitive with other rule extraction methods and effectively handles time series.

comment: - Fixed several typos - Related work have been expanded - Discussion of computational results has been improved for clearness

♻ ☆ Federated Stochastic Approximation under Markov Noise and Heterogeneity: Applications in Reinforcement Learning ICML 2022

Sajad Khodadadian, Pranay Sharma, Gauri Joshi, Siva Theja Maguluri

Since reinforcement learning algorithms are notoriously data-intensive, the task of sampling observations from the environment is usually split across multiple agents. However, transferring these observations from the agents to a central location can be prohibitively expensive in terms of communication cost, and it can also compromise the privacy of each agent's local behavior policy. Federated reinforcement learning is a framework in which $N$ agents collaboratively learn a global model, without sharing their individual data and policies. This global model is the unique fixed point of the average of $N$ local operators, corresponding to the $N$ agents. Each agent maintains a local copy of the global model and updates it using locally sampled data. In this paper, we show that by careful collaboration of the agents in solving this joint fixed point problem, we can find the global model $N$ times faster, also known as linear speedup. We first propose a general framework for federated stochastic approximation with Markovian noise and heterogeneity, showing linear speedup in convergence. We then apply this framework to federated reinforcement learning algorithms, examining the convergence of federated on-policy TD, off-policy TD, and $Q$-learning.

comment: 80 pages, 0 figure, accepted to ICML 2022 for long presentation

♻ ☆ Comparing Differentiable and Dynamic Ray Tracing: Introducing the Multipath Lifetime Map

Jérome Eertmans, Enrico Maria Vittuci, Vittorio Degli Esposti, Laurent Jacques, Claude Oestges

With the increasing presence of dynamic scenarios, such as Vehicle-to-Vehicle communications, radio propagation modeling tools must adapt to the rapidly changing nature of the radio channel. Recently, both Differentiable and Dynamic Ray Tracing frameworks have emerged to address these challenges. However, there is often confusion about how these approaches differ and which one should be used in specific contexts. In this paper, we provide an overview of these two techniques and a comparative analysis against two state-of-the-art tools: 3DSCAT from UniBo and Sionna from NVIDIA. To provide a more precise characterization of the scope of these methods, we introduce a novel simulation-based metric, the Multipath Lifetime Map, which enables the evaluation of spatial and temporal coherence in radio channels only based on the geometrical description of the environment. Finally, our metrics are evaluated on a classic urban street canyon scenario, yielding similar results to those obtained from measurement campaigns.

comment: 5 pages, 5 figures, 1 table, submitted to EuCAP 2025

♻ ☆ FoundTS: Comprehensive and Unified Benchmarking of Foundation Models for Time Series Forecasting

Zhe Li, Xiangfei Qiu, Peng Chen, Yihang Wang, Hanyin Cheng, Yang Shu, Jilin Hu, Chenjuan Guo, Aoying Zhou, Qingsong Wen, Christian S. Jensen, Bin Yang

Time Series Forecasting (TSF) is key functionality in numerous fields, including in finance, weather services, and energy management. While TSF methods are emerging these days, many of them require domain-specific data collection and model training and struggle with poor generalization performance on new domains. Foundation models aim to overcome this limitation. Pre-trained on large-scale language or time series data, they exhibit promising inferencing capabilities in new or unseen data. This has spurred a surge in new TSF foundation models. We propose a new benchmark, FoundTS, to enable thorough and fair evaluation and comparison of such models. FoundTS covers a variety of TSF foundation models, including those based on large language models and those pretrained on time series. Next, FoundTS supports different forecasting strategies, including zero-shot, few-shot, and full-shot, thereby facilitating more thorough evaluations. Finally, FoundTS offers a pipeline that standardizes evaluation processes such as dataset splitting, loading, normalization, and few-shot sampling, thereby facilitating fair evaluations. Building on this, we report on an extensive evaluation of TSF foundation models on a broad range of datasets from diverse domains and with different statistical characteristics. Specifically, we identify pros and cons and inherent limitations of existing foundation models, and we identify directions for future model design. We make our code and datasets available at https://anonymous.4open.science/r/FoundTS-C2B0.

♻ ☆ Deconstructing The Ethics of Large Language Models from Long-standing Issues to New-emerging Dilemmas: A Survey

Chengyuan Deng, Yiqun Duan, Xin Jin, Heng Chang, Yijun Tian, Han Liu, Yichen Wang, Kuofeng Gao, Henry Peng Zou, Yiqiao Jin, Yijia Xiao, Shenghao Wu, Zongxing Xie, Weimin Lyu, Sihong He, Lu Cheng, Haohan Wang, Jun Zhuang

Large Language Models (LLMs) have achieved unparalleled success across diverse language modeling tasks in recent years. However, this progress has also intensified ethical concerns, impacting the deployment of LLMs in everyday contexts. This paper provides a comprehensive survey of ethical challenges associated with LLMs, from longstanding issues such as copyright infringement, systematic bias, and data privacy, to emerging problems like truthfulness and social norms. We critically analyze existing research aimed at understanding, examining, and mitigating these ethical risks. Our survey underscores integrating ethical standards and societal values into the development of LLMs, thereby guiding the development of responsible and ethically aligned language models.

♻ ☆ Wireless Human-Machine Collaboration in Industry 5.0

Gaoyang Pang, Wanchun Liu, Dusit Niyato, Daniel Quevedo, Branka Vucetic, Yonghui Li

Wireless Human-Machine Collaboration (WHMC) represents a critical advancement for Industry 5.0, enabling seamless interaction between humans and machines across geographically distributed systems. As the WHMC systems become increasingly important for achieving complex collaborative control tasks, ensuring their stability is essential for practical deployment and long-term operation. Stability analysis certifies how the closed-loop system will behave under model randomness, which is essential for systems operating with wireless communications. However, the fundamental stability analysis of the WHMC systems remains an unexplored challenge due to the intricate interplay between the stochastic nature of wireless communications, dynamic human operations, and the inherent complexities of control system dynamics. This paper establishes a fundamental WHMC model incorporating dual wireless loops for machine and human control. Our framework accounts for practical factors such as short-packet transmissions, fading channels, and advanced HARQ schemes. We model human control lag as a Markov process, which is crucial for capturing the stochastic nature of human interactions. Building on this model, we propose a stochastic cycle-cost-based approach to derive a stability condition for the WHMC system, expressed in terms of wireless channel statistics, human dynamics, and control parameters. Our findings are validated through extensive numerical simulations and a proof-of-concept experiment, where we developed and tested a novel wireless collaborative cart-pole control system. The results confirm the effectiveness of our approach and provide a robust framework for future research on WHMC systems in more complex environments.

comment: This work has been submitted to the IEEE for possible publication

♻ ☆ A multi-task deep learning approach for lane-level pavement performance prediction with segment-level data

Bo Wang, Wenbo Zhang, Yunpeng LI

The elaborate pavement performance prediction is an important premise of implementing preventive maintenance. Our survey reveals that in practice, the pavement performance is usually measured at segment-level, where an unique performance value is obtained for all lanes within one segment of 1km length. It still lacks more elaborate performance analysis at lane-level due to costly data collection and difficulty in prediction modeling. Therefore, this study developed a multi-task deep learning approach to predict the lane-level pavement performance with a large amount of historical segment-level performance measurement data. The unified prediction framework can effectively address inherent correlation and differences across lanes. In specific, the prediction framework firstly employed an Long Short-Term Memory (LSTM) layer to capture the segment-level pavement deterioration pattern. Then multiple task-specific LSTM layers were designed based on number of lanes to capture lane-level differences in pavement performance. Finally, we concatenated multiple task-specific LSTM outputs with auxiliary features for further training and obtained the lane-level predictions after fully connected layer. The aforementioned prediction framework was validated with a real case in China. It revealed a better model performance regardless of one-way 2-lane, 3-lane, and 4-lane scenarios, all lower than 10% in terms of mean absolute percentage error. The proposed prediction framework also outperforms other ensemble learning and shallow machine learning methods in almost every lane.

comment: 24 pages, 8 figures, 4 tables

♻ ☆ Reinforcement Learning for Molecular Dynamics Optimization: A Stochastic Pontryagin Maximum Principle Approach ICONIP

Chandrajit Bajaj, Minh Nguyen, Conrad Li

In this paper, we present a novel reinforcement learning framework designed to optimize molecular dynamics by focusing on the entire trajectory rather than just the final molecular configuration. Leveraging a stochastic version of Pontryagin's Maximum Principle (PMP) and Soft Actor-Critic (SAC) algorithm, our framework effectively explores non-convex molecular energy landscapes, escaping local minima to stabilize in low-energy states. Our approach operates in continuous state and action spaces without relying on labeled data, making it applicable to a wide range of molecular systems. Through extensive experimentation on six distinct molecules, including Bradykinin and Oxytocin, we demonstrate competitive performance against other unsupervised physics-based methods, such as the Greedy and NEMO-based algorithms. Our method's adaptability and focus on dynamic trajectory optimization make it suitable for applications in areas such as drug discovery and molecular design.

comment: Accepted to the International Conference on Neural Information Processing (ICONIP) 2024. To be published in Springer-Nature Communications in Computer and Information Science (CCIS) Series

♻ ☆ Self-Supervised Learning of Disentangled Representations for Multivariate Time-Series NeurIPS

Ching Chang, Chiao-Tung Chan, Wei-Yao Wang, Wen-Chih Peng, Tien-Fu Chen

Multivariate time-series data in fields like healthcare and industry are informative but challenging due to high dimensionality and lack of labels. Recent self-supervised learning methods excel in learning rich representations without labels but struggle with disentangled embeddings and inductive bias issues like transformation-invariance. To address these challenges, we introduce TimeDRL, a framework for multivariate time-series representation learning with dual-level disentangled embeddings. TimeDRL features: (i) disentangled timestamp-level and instance-level embeddings using a [CLS] token strategy; (ii) timestamp-predictive and instance-contrastive tasks for representation learning; and (iii) avoidance of augmentation methods to eliminate inductive biases. Experiments on forecasting and classification datasets show TimeDRL outperforms existing methods, with further validation in semi-supervised settings with limited labeled data.

comment: This submission has been withdrawn to avoid duplication with a full version of the paper that is already available in another arXiv entry (arXiv:2410.12606). The withdrawn version was a short format prepared for a NeurIPS workshop and is no longer necessary as a separate arXiv submission

♻ ☆ TinyAgent: Function Calling at the Edge EMNLP 2024

Lutfi Eren Erdogan, Nicholas Lee, Siddharth Jha, Sehoon Kim, Ryan Tabrizi, Suhong Moon, Coleman Hooper, Gopala Anumanchipalli, Kurt Keutzer, Amir Gholami

Recent large language models (LLMs) have enabled the development of advanced agentic systems that can integrate various tools and APIs to fulfill user queries through function calling. However, the deployment of these LLMs on the edge has not been explored since they typically require cloud-based infrastructure due to their substantial model size and computational demands. To this end, we present TinyAgent, an end-to-end framework for training and deploying task-specific small language model agents capable of function calling for driving agentic systems at the edge. We first show how to enable accurate function calling for open-source models via the LLMCompiler framework. We then systematically curate a high-quality dataset for function calling, which we use to fine-tune two small language models, TinyAgent-1.1B and 7B. For efficient inference, we introduce a novel tool retrieval method to reduce the input prompt length and utilize quantization to further accelerate the inference speed. As a driving application, we demonstrate a local Siri-like system for Apple's MacBook that can execute user commands through text or voice input. Our results show that our models can achieve, and even surpass, the function-calling capabilities of larger models like GPT-4-Turbo, while being fully deployed at the edge. We open-source our dataset, models, and installable package and provide a demo video for our MacBook assistant agent.

comment: EMNLP 2024 Demo

♻ ☆ Exploring Singularities in point clouds with the graph Laplacian: An explicit approach

Martin Andersson, Benny Avelin

We develop theory and methods that use the graph Laplacian to analyze the geometry of the underlying manifold of datasets. Our theory provides theoretical guarantees and explicit bounds on the functional forms of the graph Laplacian when it acts on functions defined close to singularities of the underlying manifold. We use these explicit bounds to develop tests for singularities and propose methods that can be used to estimate geometric properties of singularities in the datasets.

comment: 27 pages, 12 figures

♻ ☆ A Rainbow in Deep Network Black Boxes

Florentin Guth, Brice Ménard, Gaspar Rochette, Stéphane Mallat

A central question in deep learning is to understand the functions learned by deep networks. What is their approximation class? Do the learned weights and representations depend on initialization? Previous empirical work has evidenced that kernels defined by network activations are similar across initializations. For shallow networks, this has been theoretically studied with random feature models, but an extension to deep networks has remained elusive. Here, we provide a deep extension of such random feature models, which we call the rainbow model. We prove that rainbow networks define deterministic (hierarchical) kernels in the infinite-width limit. The resulting functions thus belong to a data-dependent RKHS which does not depend on the weight randomness. We also verify numerically our modeling assumptions on deep CNNs trained on image classification tasks, and show that the trained networks approximately satisfy the rainbow hypothesis. In particular, rainbow networks sampled from the corresponding random feature model achieve similar performance as the trained networks. Our results highlight the central role played by the covariances of network weights at each layer, which are observed to be low-rank as a result of feature learning.

comment: 59 pages, 10 figures. To appear at JMLR

♻ ☆ Concentration of the Langevin Algorithm's Stationary Distribution

Jason M. Altschuler, Kunal Talwar

A canonical algorithm for log-concave sampling is the Langevin Algorithm, aka the Langevin Diffusion run with some discretization stepsize $\eta > 0$. This discretization leads the Langevin Algorithm to have a stationary distribution $\pi_{\eta}$ which differs from the stationary distribution $\pi$ of the Langevin Diffusion, and it is an important challenge to understand whether the well-known properties of $\pi$ extend to $\pi_{\eta}$. In particular, while concentration properties such as isoperimetry and rapidly decaying tails are classically known for $\pi$, the analogous properties for $\pi_{\eta}$ are open questions with algorithmic implications. This note provides a first step in this direction by establishing concentration results for $\pi_{\eta}$ that mirror classical results for $\pi$. Specifically, we show that for any nontrivial stepsize $\eta > 0$, $\pi_{\eta}$ is sub-exponential (respectively, sub-Gaussian) when the potential is convex (respectively, strongly convex). Moreover, the concentration bounds we show are essentially tight. We also show that these concentration bounds extend to all iterates along the trajectory of the Langevin Algorithm, and to inexact implementations which use sub-Gaussian estimates of the gradient. Key to our analysis is the use of a rotation-invariant moment generating function (aka Bessel function) to study the stationary dynamics of the Langevin Algorithm. This technique may be of independent interest because it enables directly analyzing the discrete-time stationary distribution $\pi_{\eta}$ without going through the continuous-time stationary distribution $\pi$ as an intermediary.

comment: Added Section 6 (extensions to concentration of the trajectory and inexact gradients)

♻ ☆ FAMOUS: Flexible Accelerator for the Attention Mechanism of Transformer on UltraScale+ FPGAs

Ehsan Kabir, Md. Arafat Kabir, Austin R. J. Downey, Jason D. Bakos, David Andrews, Miaoqing Huang

Transformer neural networks (TNNs) are being applied across a widening range of application domains, including natural language processing (NLP), machine translation, and computer vision (CV). Their popularity is largely attributed to the exceptional performance of their multi-head self-attention blocks when analyzing sequential data and extracting features. To date, there are limited hardware accelerators tailored for this mechanism, which is the first step before designing an accelerator for a complete model. This paper proposes \textit{FAMOUS}, a flexible hardware accelerator for dense multi-head attention (MHA) computation of TNNs on field-programmable gate arrays (FPGAs). It is optimized for high utilization of processing elements and on-chip memories to improve parallelism and reduce latency. An efficient tiling of large matrices has been employed to distribute memory and computing resources across different modules on various FPGA platforms. The design is evaluated on Xilinx Alveo U55C and U200 data center cards containing Ultrascale+ FPGAs. Experimental results are presented that show that it can attain a maximum throughput, number of parallel attention heads, embedding dimension and tile size of 328 (giga operations/second (GOPS)), 8, 768 and 64 respectively on the U55C. Furthermore, it is 3.28$\times$ and 2.6$\times$ faster than the Intel Xeon Gold 5220R CPU and NVIDIA V100 GPU respectively. It is also 1.3$\times$ faster than the fastest state-of-the-art FPGA-based accelerator.

comment: arXiv admin note: text overlap with arXiv:2409.13975

♻ ☆ FSL-Rectifier: Rectify Outliers in Few-Shot Learning via Test-Time Augmentation

Yunwei Bai, Ying Kiat Tan, Shiming Chen, Yao Shu, Tsuhan Chen

Few-shot-learning (FSL) commonly requires a model to identify images (queries) that belong to classes unseen during training, based on a few labeled samples of the new classes (support set) as reference. So far, plenty of algorithms involve training data augmentation to improve the generalization capability of FSL models, but outlier queries or support images during inference can still pose great generalization challenges. In this work, to reduce the bias caused by the outlier samples, we generate additional test-class samples by combining original samples with suitable train-class samples via a generative image combiner. Then, we obtain averaged features via an augmentor, which leads to more typical representations through the averaging. We experimentally and theoretically demonstrate the effectiveness of our method, e.g., obtaining a test accuracy improvement proportion of around 10% (e.g., from 46.86% to 53.28%) for trained FSL models. Importantly, given pretrained image combiner, our method is training-free for off-the-shelf FSL models, whose performance can be improved without extra datasets nor further training of the models themselves.

♻ ☆ AlignedKV: Reducing Memory Access of KV-Cache with Precision-Aligned Quantization

Yifan Tan, Haoze Wang, Chao Yan, Yangdong Deng

Model quantization has become a crucial technique to address the issues of large memory consumption and long inference times associated with LLMs. Mixed-precision quantization, which distinguishes between important and unimportant parameters, stands out among numerous quantization schemes as it achieves a balance between precision and compression rate. However, existing approaches can only identify important parameters through qualitative analysis and manual experiments without quantitatively analyzing how their importance is determined. We propose a new criterion, so-called 'precision alignment', to build a quantitative framework to holistically evaluate the importance of parameters in mixed-precision quantization. Our observations on floating point addition under various real-world scenarios suggest that two addends should have identical precision, otherwise the information in the higher-precision number will be wasted. Such an observation offers an essential principle to determine the precision of each parameter in matrix multiplication operation. As the first step towards applying the above discovery to large model inference, we develop a dynamic KV-Cache quantization technique to effectively reduce memory access latency. Different from existing quantization approaches that focus on memory saving, this work directly aims to accelerate LLM inference through quantifying floating numbers. The proposed technique attains a 25% saving of memory access and delivers up to 1.3x speedup in the computation of attention in the decoding phase of LLM, with almost no loss of precision.

♻ ☆ An Asymptotically Optimal Algorithm for the Convex Hull Membership Problem

Gang Qiao, Ambuj Tewari

We study the convex hull membership (CHM) problem in the pure exploration setting where one aims to efficiently and accurately determine if a given point lies in the convex hull of means of a finite set of distributions. We give a complete characterization of the sample complexity of the CHM problem in the one-dimensional case. We present the first asymptotically optimal algorithm called Thompson-CHM, whose modular design consists of a stopping rule and a sampling rule. In addition, we extend the algorithm to settings that generalize several important problems in the multi-armed bandit literature. Furthermore, we discuss the extension of Thompson-CHM to higher dimensions. Finally, we provide numerical experiments to demonstrate the empirical behavior of the algorithm matches our theoretical results for realistic time horizons.

♻ ☆ Diffusion-TS: Interpretable Diffusion for General Time Series Generation

Xinyu Yuan, Yan Qiao

Denoising diffusion probabilistic models (DDPMs) are becoming the leading paradigm for generative models. It has recently shown breakthroughs in audio synthesis, time series imputation and forecasting. In this paper, we propose Diffusion-TS, a novel diffusion-based framework that generates multivariate time series samples of high quality by using an encoder-decoder transformer with disentangled temporal representations, in which the decomposition technique guides Diffusion-TS to capture the semantic meaning of time series while transformers mine detailed sequential information from the noisy model input. Different from existing diffusion-based approaches, we train the model to directly reconstruct the sample instead of the noise in each diffusion step, combining a Fourier-based loss term. Diffusion-TS is expected to generate time series satisfying both interpretablity and realness. In addition, it is shown that the proposed Diffusion-TS can be easily extended to conditional generation tasks, such as forecasting and imputation, without any model changes. This also motivates us to further explore the performance of Diffusion-TS under irregular settings. Finally, through qualitative and quantitative experiments, results show that Diffusion-TS achieves the state-of-the-art results on various realistic analyses of time series.

♻ ☆ Privacy-Preserving Community Detection for Locally Distributed Multiple Networks

Xiao Guo, Xiang Li, Xiangyu Chang, Shujie Ma

Modern multi-layer networks are commonly stored and analyzed in a local and distributed fashion because of the privacy, ownership, and communication costs. The literature on the model-based statistical methods for community detection based on these data is still limited. This paper proposes a new method for consensus community detection and estimation in a multi-layer stochastic block model using locally stored and computed network data with privacy protection. A novel algorithm named privacy-preserving Distributed Spectral Clustering (ppDSC) is developed. To preserve the edges' privacy, we adopt the randomized response (RR) mechanism to perturb the network edges, which satisfies the strong notion of differential privacy. The ppDSC algorithm is performed on the squared RR-perturbed adjacency matrices to prevent possible cancellation of communities among different layers. To remove the bias incurred by RR and the squared network matrices, we develop a two-step bias-adjustment procedure. Then we perform eigen-decomposition on the debiased matrices, aggregation of the local eigenvectors using an orthogonal Procrustes transformation, and k-means clustering. We provide theoretical analysis on the statistical errors of ppDSC in terms of eigen-vector estimation. In addition, the blessings and curses of network heterogeneity are well-explained by our bounds.

♻ ☆ Diffusion Models as Constrained Samplers for Optimization with Unknown Constraints

Lingkai Kong, Yuanqi Du, Wenhao Mu, Kirill Neklyudov, Valentin De Bortoli, Dongxia Wu, Haorui Wang, Aaron Ferber, Yi-An Ma, Carla P. Gomes, Chao Zhang

Addressing real-world optimization problems becomes particularly challenging when analytic objective functions or constraints are unavailable. While numerous studies have addressed the issue of unknown objectives, limited research has focused on scenarios where feasibility constraints are not given explicitly. Overlooking these constraints can lead to spurious solutions that are unrealistic in practice. To deal with such unknown constraints, we propose to perform optimization within the data manifold using diffusion models. To constrain the optimization process to the data manifold, we reformulate the original optimization problem as a sampling problem from the product of the Boltzmann distribution defined by the objective function and the data distribution learned by the diffusion model. Depending on the differentiability of the objective function, we propose two different sampling methods. For differentiable objectives, we propose a two-stage framework that begins with a guided diffusion process for warm-up, followed by a Langevin dynamics stage for further correction. For non-differentiable objectives, we propose an iterative importance sampling strategy using the diffusion model as the proposal distribution. Comprehensive experiments on a synthetic dataset, six real-world black-box optimization datasets, and a multi-objective molecule optimization dataset show that our method achieves better or comparable performance with previous state-of-the-art baselines.

♻ ☆ Open-World Continual Learning: Unifying Novelty Detection and Continual Learning

Gyuhak Kim, Changnan Xiao, Tatsuya Konishi, Zixuan Ke, Bing Liu

As AI agents are increasingly used in the real open world with unknowns or novelties, they need the ability to (1) recognize objects that (a) they have learned before and (b) detect items that they have never seen or learned, and (2) learn the new items incrementally to become more and more knowledgeable and powerful. (1) is called novelty detection or out-of-distribution (OOD) detection and (2) is called class incremental learning (CIL), which is a setting of continual learning (CL). In existing research, OOD detection and CIL are regarded as two completely different problems. This paper first provides a theoretical proof that good OOD detection for each task within the set of learned tasks (called closed-world OOD detection) is necessary for successful CIL. We show this by decomposing CIL into two sub-problems: within-task prediction (WP) and task-id prediction (TP), and proving that TP is correlated with closed-world OOD detection. The key theoretical result is that regardless of whether WP and OOD detection (or TP) are defined explicitly or implicitly by a CIL algorithm, good WP and good closed-world OOD detection are necessary and sufficient conditions for good CIL, which unifies novelty or OOD detection and continual learning (CIL, in particular). We call this traditional CIL the closed-world CIL as it does not detect future OOD data in the open world. The paper then proves that the theory can be generalized or extended to open-world CIL, which is the proposed open-world continual learning, that can perform CIL in the open world and detect future or open-world OOD data. Based on the theoretical results, new CIL methods are also designed, which outperform strong baselines in CIL accuracy and in continual OOD detection by a large margin.

comment: To appear in Artificial Intelligence Journal. arXiv admin note: substantial text overlap with arXiv:2211.02633

♻ ☆ Causally Testing Gender Bias in LLMs: A Case Study on Occupational Bias

Yuen Chen, Vethavikashini Chithrra Raghuram, Justus Mattern, Rada Mihalcea, Zhijing Jin

Generated texts from large language models (LLMs) have been shown to exhibit a variety of harmful, human-like biases against various demographics. These findings motivate research efforts aiming to understand and measure such effects. This paper introduces a causal formulation for bias measurement in generative language models. Based on this theoretical foundation, we outline a list of desiderata for designing robust bias benchmarks. We then propose a benchmark called OccuGender, with a bias-measuring procedure to investigate occupational gender bias. We test several state-of-the-art open-source LLMs on OccuGender, including Llama, Mistral, and their instruction-tuned versions. The results show that these models exhibit substantial occupational gender bias. Lastly, we discuss prompting strategies for bias mitigation and an extension of our causal formulation to illustrate the generalizability of our framework. Our code and data https://github.com/chenyuen0103/gender-bias.

♻ ☆ Adversarial Consistency and the Uniqueness of the Adversarial Bayes Classifier

Natalie S. Frank

Minimizing an adversarial surrogate risk is a common technique for learning robust classifiers. Prior work showed that convex surrogate losses are not statistically consistent in the adversarial context -- or in other words, a minimizing sequence of the adversarial surrogate risk will not necessarily minimize the adversarial classification error. We connect the consistency of adversarial surrogate losses to properties of minimizers to the adversarial classification risk, known as adversarial Bayes classifiers. Specifically, under reasonable distributional assumptions, a convex surrogate loss is statistically consistent for adversarial learning iff the adversarial Bayes classifier satisfies a certain notion of uniqueness.

comment: 2 figures, 20 pages, v2: fixed typos, v3: improved organization of paper and added figures

♻ ☆ CinePile: A Long Video Question Answering Dataset and Benchmark

Ruchit Rawal, Khalid Saifullah, Miquel Farré, Ronen Basri, David Jacobs, Gowthami Somepalli, Tom Goldstein

Current datasets for long-form video understanding often fall short of providing genuine long-form comprehension challenges, as many tasks derived from these datasets can be successfully tackled by analyzing just one or a few random frames from a video. To address this issue, we present a novel dataset and benchmark, CinePile, specifically designed for authentic long-form video understanding. This paper details our innovative approach for creating a question-answer dataset, utilizing advanced LLMs with human-in-the-loop and building upon human-generated raw data. Our comprehensive dataset comprises 305,000 multiple-choice questions (MCQs), covering various visual and multimodal aspects, including temporal comprehension, understanding human-object interactions, and reasoning about events or actions within a scene. Additionally, we fine-tuned open-source Video-LLMs on the training split and evaluated both open-source and proprietary video-centric LLMs on the test split of our dataset. The findings indicate that although current models underperform compared to humans, fine-tuning these models can lead to significant improvements in their performance.

comment: Project page with all the artifacts - https://ruchitrawal.github.io/cinepile/. Updated version with adversarial refinement pipeline and more model evaluations

♻ ☆ Learning Personalized Scoping for Graph Neural Networks under Heterophily

Gangda Deng, Hongkuan Zhou, Rajgopal Kannan, Viktor Prasanna

Heterophilous graphs, where dissimilar nodes tend to connect, pose a challenge for graph neural networks (GNNs) as their superior performance typically comes from aggregating homophilous information. Increasing the GNN depth can expand the scope (i.e., receptive field), potentially finding homophily from the higher-order neighborhoods. However, uniformly expanding the scope results in subpar performance since real-world web graphs often exhibit homophily disparity between nodes. An ideal way is personalized scopes, allowing nodes to have varying scope sizes. Existing methods typically add node-adaptive weights for each hop. Although expressive, they inevitably suffer from severe overfitting. To address this issue, we formalize personalized scoping as a separate scope classification problem that overcomes GNN overfitting in node classification. Specifically, we predict the optimal GNN depth for each node. Our theoretical and empirical analysis suggests that accurately predicting the depth can significantly enhance generalization. We further propose Adaptive Scope (AS), a lightweight approach that only participates in GNN inference. AS encodes structural patterns and predicts the depth to select the best model for each node's prediction. Experimental results show that AS is highly flexible with various GNN architectures across a wide range of datasets while significantly improving accuracy.

♻ ☆ DEPT: Decoupled Embeddings for Pre-training Language Models

Alex Iacob, Lorenzo Sani, Meghdad Kurmanji, William F. Shen, Xinchi Qiu, Dongqi Cai, Yan Gao, Nicholas D. Lane

Language model pre-training benefits from diverse data to enhance performance across domains and languages. However, training on such heterogeneous corpora requires extensive and costly efforts. Since these data sources vary lexically, syntactically, and semantically, they cause negative interference or the ``curse of multilinguality''. We propose a novel pre-training framework to alleviate this curse. Our method, DEPT, decouples embeddings from the transformer body while simultaneously training the latter in multiple contexts. DEPT enables training without a shared global vocabulary and: (1) can train robustly and effectively under significant data heterogeneity, (2) reduces token embedding parameters by up to 80% and the communication costs by 675x for billion-scale models, (3) enhances model generalization and plasticity in adapting to new languages and domains, and (4) permits training with custom optimized vocabularies per data source. We demonstrate DEPT's potential via the first vocabulary-agnostic federated multilingual pre-training of a 1.3 billion-parameter model, limiting its embedding size to 102.4 million instead of 512 million.

♻ ☆ Kernel PCA for Out-of-Distribution Detection NeurIPS 2024

Kun Fang, Qinghua Tao, Kexin Lv, Mingzhen He, Xiaolin Huang, Jie Yang

Out-of-Distribution (OoD) detection is vital for the reliability of Deep Neural Networks (DNNs). Existing works have shown the insufficiency of Principal Component Analysis (PCA) straightforwardly applied on the features of DNNs in detecting OoD data from In-Distribution (InD) data. The failure of PCA suggests that the network features residing in OoD and InD are not well separated by simply proceeding in a linear subspace, which instead can be resolved through proper non-linear mappings. In this work, we leverage the framework of Kernel PCA (KPCA) for OoD detection, and seek suitable non-linear kernels that advocate the separability between InD and OoD data in the subspace spanned by the principal components. Besides, explicit feature mappings induced from the devoted task-specific kernels are adopted so that the KPCA reconstruction error for new test samples can be efficiently obtained with large-scale data. Extensive theoretical and empirical results on multiple OoD data sets and network structures verify the superiority of our KPCA detector in efficiency and efficacy with state-of-the-art detection performance.

comment: Accepted by NeurIPS 2024

♻ ☆ LW-FedSSL: Resource-efficient Layer-wise Federated Self-supervised Learning

Ye Lin Tun, Chu Myaet Thwal, Le Quang Huy, Minh N. H. Nguyen, Choong Seon Hong

Many studies integrate federated learning (FL) with self-supervised learning (SSL) to take advantage of raw data distributed across edge devices. However, edge devices often struggle with high computation and communication costs imposed by SSL and FL algorithms. To tackle this hindrance, we propose LW-FedSSL, a layer-wise federated self-supervised learning approach that allows edge devices to incrementally train a single layer of the model at a time. We introduce server-side calibration and representation alignment mechanisms to ensure LW-FedSSL delivers performance on par with conventional federated self-supervised learning (FedSSL) while significantly lowering resource demands. In a pure layer-wise training scheme, training one layer at a time may limit effective interaction between different layers of the model. The server-side calibration mechanism takes advantage of the resource-rich FL server to ensure smooth collaboration between different layers of the global model. During local training, the representation alignment mechanism encourages closeness between representations of local models and those of the global model, thereby preserving the layer cohesion established by server-side calibration. With the proposed mechanisms, LW-FedSSL achieves a $3.3 \times$ reduction in memory usage, $2.1 \times$ fewer computational operations (FLOPs), and a $3.2 \times$ lower communication cost while maintaining the same level of performance as its end-to-end training counterpart. Additionally, we explore a progressive training strategy called Prog-FedSSL, which matches end-to-end training in memory requirements but offers a $1.8 \times$ reduction in FLOPs and communication costs. Although Prog-FedSSL is not as resource-efficient as LW-FedSSL, its performance improvements make it a suitable candidate for FL environments with more lenient resource constraints.

♻ ☆ A Statistical View of Column Subset Selection

Anav Sood, Trevor Hastie

We consider the problem of selecting a small subset of representative variables from a large dataset. In the computer science literature, this dimensionality reduction problem is typically formalized as Column Subset Selection (CSS). Meanwhile, the typical statistical formalization is to find an information-maximizing set of Principal Variables. This paper shows that these two approaches are equivalent, and moreover, both can be viewed as maximum likelihood estimation within a certain semi-parametric model. Within this model, we establish suitable conditions under which the CSS estimate is consistent in high dimensions, specifically in the proportional asymptotic regime where the number of variables over the sample size converges to a constant. Using these connections, we show how to efficiently (1) perform CSS using only summary statistics from the original dataset; (2) perform CSS in the presence of missing and/or censored data; and (3) select the subset size for CSS in a hypothesis testing framework.

♻ ☆ Resource-Efficient Federated Multimodal Learning via Layer-wise and Progressive Training

Ye Lin Tun, Chu Myaet Thwal, Minh N. H. Nguyen, Choong Seon Hong

Combining different data modalities enables deep neural networks to tackle complex tasks more effectively, making multimodal learning increasingly popular. To harness multimodal data closer to end users, it is essential to integrate multimodal learning with privacy-preserving approaches like federated learning (FL). However, compared to conventional unimodal learning, multimodal setting requires dedicated encoders for each modality, resulting in larger and more complex models. Training these models requires significant resources, presenting a substantial challenge for FL clients operating with limited computation and communication resources. To address these challenges, we introduce LW-FedMML, a layer-wise federated multimodal learning approach which decomposes the training process into multiple stages. Each stage focuses on training only a portion of the model, thereby significantly reducing the memory and computational requirements. Moreover, FL clients only need to exchange the trained model portion with the central server, lowering the resulting communication cost. We conduct extensive experiments across various FL and multimodal learning settings to validate the effectiveness of our proposed method. The results demonstrate that LW-FedMML can compete with conventional end-to-end federated multimodal learning (FedMML) while significantly reducing the resource burden on FL clients. Specifically, LW-FedMML reduces memory usage by up to $2.7\times$, computational operations (FLOPs) by $2.4\times$, and total communication cost by $2.3\times$. We also explore a progressive training approach called Prog-FedMML. While it offers lesser resource efficiency than LW-FedMML, Prog-FedMML has the potential to surpass the performance of end-to-end FedMML, making it a viable option for scenarios with fewer resource constraints.

♻ ☆ Exogenous Matching: Learning Good Proposals for Tractable Counterfactual Estimation

Yikang Chen, Dehui du, Lili Tian

We propose an importance sampling method for tractable and efficient estimation of counterfactual expressions in general settings, named Exogenous Matching. By minimizing a common upper bound of counterfactual estimators, we transform the variance minimization problem into a conditional distribution learning problem, enabling its integration with existing conditional distribution modeling approaches. We validate the theoretical results through experiments under various types and settings of Structural Causal Models (SCMs) and demonstrate the outperformance on counterfactual estimation tasks compared to other existing importance sampling methods. We also explore the impact of injecting structural prior knowledge (counterfactual Markov boundaries) on the results. Finally, we apply this method to identifiable proxy SCMs and demonstrate the unbiasedness of the estimates, empirically illustrating the applicability of the method to practical scenarios.

comment: 51 pages, 15 figures

♻ ☆ Balancing Label Quantity and Quality for Scalable Elicitation

Alex Mallen, Nora Belrose

Scalable oversight studies methods of training and evaluating AI systems in domains where human judgment is unreliable or expensive, such as scientific research and software engineering in complex codebases. Most work in this area has focused on methods of improving the quality of labels. Recent work by Burns et al. (2023) considers the complementary problem of training models with low-quality labels, finding that large pretrained models often have an inductive bias towards producing correct answers. In practice, however, neither label quantity nor quality is fixed: practitioners face a quantity-quality tradeoff. In this paper, we explore the microeconomics of the quantity-quality tradeoff on binary NLP classification tasks used in Burns et al. (2023). While sample-efficient learning has been studied extensively, little public research has focused on scalable elicitation: eliciting capabilities from pretrained models subject to labeling cost constraints. We find that this setting has novel dynamics caused by the tradeoff between label quantity and quality, as well as the model's existing latent capabilities. We observe three regimes of eliciting classification knowledge from pretrained models using supervised finetuning: quantity-dominant, quality-dominant, and a mixed regime involving the use of low- and high-quality data together to attain higher accuracy at a lower cost than using either alone. We explore sample-efficient elicitation methods that make use of two datasets of differing qualities, and establish a Pareto frontier of scalable elicitation methods that optimally trade off labeling cost and classifier performance. We find that the accuracy of supervised fine-tuning can be improved by up to 5 percentage points at a fixed labeling budget by adding a few-shot prompt to make use of the model's existing knowledge of the task.

♻ ☆ TrojanForge: Generating Adversarial Hardware Trojan Examples with Reinforcement Learning

Amin Sarihi, Peter Jamieson, Ahmad Patooghy, Abdel-Hameed A. Badawy

The Hardware Trojan (HT) problem can be thought of as a continuous game between attackers and defenders, each striving to outsmart the other by leveraging any available means for an advantage. Machine Learning (ML) has recently played a key role in advancing HT research. Various novel techniques, such as Reinforcement Learning (RL) and Graph Neural Networks (GNNs), have shown HT insertion and detection capabilities. HT insertion with ML techniques, specifically, has seen a spike in research activity due to the shortcomings of conventional HT benchmarks and the inherent human design bias that occurs when we create them. This work continues this innovation by presenting a tool called TrojanForge, capable of generating HT adversarial examples that defeat HT detectors; demonstrating the capabilities of GAN-like adversarial tools for automatic HT insertion. We introduce an RL environment where the RL insertion agent interacts with HT detectors in an insertion-detection loop where the agent collects rewards based on its success in bypassing HT detectors. Our results show that this process helps inserted HTs evade various HT detectors, achieving high attack success percentages. This tool provides insight into why HT insertion fails in some instances and how we can leverage this knowledge in defense.

♻ ☆ Is Prior-Free Black-Box Non-Stationary Reinforcement Learning Feasible?

Argyrios Gerogiannis, Yu-Han Huang, Venugopal V. Veeravalli

We study the problem of Non-Stationary Reinforcement Learning (NS-RL) without prior knowledge about the system's non-stationarity. A state-of-the-art, black-box algorithm, known as MASTER, is considered, with a focus on identifying the conditions under which it can achieve its stated goals. Specifically, we prove that MASTER's non-stationarity detection mechanism is not triggered for practical choices of horizon, leading to performance akin to a random restarting algorithm. Moreover, we show that the regret bound for MASTER, while being order optimal, stays above the worst-case linear regret until unreasonably large values of the horizon. To validate these observations, MASTER is tested for the special case of piecewise stationary multi-armed bandits, along with methods that employ random restarting, and others that use quickest change detection to restart. A simple, order optimal random restarting algorithm, that has prior knowledge of the non-stationarity is proposed as a baseline. The behavior of the MASTER algorithm is validated in simulations, and it is shown that methods employing quickest change detection are more robust and consistently outperform MASTER and other random restarting approaches.

comment: Corrected minor typos in the proof of Theorem 2 on pages 25 and 26

♻ ☆ GRAG: Graph Retrieval-Augmented Generation

Yuntong Hu, Zhihan Lei, Zheng Zhang, Bo Pan, Chen Ling, Liang Zhao

Naive Retrieval-Augmented Generation (RAG) focuses on individual documents during retrieval and, as a result, falls short in handling networked documents which are very popular in many applications such as citation graphs, social media, and knowledge graphs. To overcome this limitation, we introduce Graph Retrieval-Augmented Generation (GRAG), which tackles the fundamental challenges in retrieving textual subgraphs and integrating the joint textual and topological information into Large Language Models (LLMs) to enhance its generation. To enable efficient textual subgraph retrieval, we propose a novel divide-and-conquer strategy that retrieves the optimal subgraph structure in linear time. To achieve graph context-aware generation, incorporate textual graphs into LLMs through two complementary views-the text view and the graph view-enabling LLMs to more effectively comprehend and utilize the graph context. Extensive experiments on graph reasoning benchmarks demonstrate that in scenarios requiring multi-hop reasoning on textual graphs, our GRAG approach significantly outperforms current state-of-the-art RAG methods.

comment: 13 pages, 5 figures

♻ ☆ Stepwise Alignment for Constrained Language Model Policy Optimization NeurIPS 2024

Akifumi Wachi, Thien Q. Tran, Rei Sato, Takumi Tanabe, Youhei Akimoto

Safety and trustworthiness are indispensable requirements for real-world applications of AI systems using large language models (LLMs). This paper formulates human value alignment as an optimization problem of the language model policy to maximize reward under a safety constraint, and then proposes an algorithm, Stepwise Alignment for Constrained Policy Optimization (SACPO). One key idea behind SACPO, supported by theory, is that the optimal policy incorporating reward and safety can be directly obtained from a reward-aligned policy. Building on this key idea, SACPO aligns LLMs step-wise with each metric while leveraging simple yet powerful alignment algorithms such as direct preference optimization (DPO). SACPO offers several advantages, including simplicity, stability, computational efficiency, and flexibility of algorithms and datasets. Under mild assumptions, our theoretical analysis provides the upper bounds on optimality and safety constraint violation. Our experimental results show that SACPO can fine-tune Alpaca-7B better than the state-of-the-art method in terms of both helpfulness and harmlessness.

comment: Accepted at NeurIPS 2024. Code and models are available at https://github.com/line/sacpo

♻ ☆ Utilizing Large Language Models in An Iterative Paradigm with Domain Feedback for Molecule Optimization

Khiem Le, Nitesh V. Chawla

Molecule optimization is a critical task in drug discovery to optimize desired properties of a given molecule through chemical modification. Despite Large Language Models (LLMs) holding the potential to efficiently simulate this task by using natural language to direct the optimization, straightforwardly utilizing shows limited performance. In this work, we facilitate utilizing LLMs in an iterative paradigm by proposing a simple yet highly effective domain feedback provider, namely $\text{Re}^2$DF. In detail, $\text{Re}^2$DF harnesses an external toolkit, RDKit, to handle the molecule hallucination, if the modified molecule is chemically invalid. Otherwise, its desired properties are computed and compared to the original one, establishing reliable domain feedback with correct direction and distance towards the objective, followed by a retrieved example, to explicitly guide the LLM to refine the modified molecule. We conduct experiments across both single- and multi-property objectives with 2 thresholds, where $\text{Re}^2$DF shows significant improvements. Particularly, for 20 single-property objectives, $\text{Re}^2$DF enhances Hit ratio by 16.95% and 20.76% under loose and strict thresholds, respectively. For 32 multi-property objectives, $\text{Re}^2$DF enhances Hit ratio by 6.04% and 5.25%.

♻ ☆ Behavior-Inspired Neural Networks for Relational Inference

Yulong Yang, Bowen Feng, Keqin Wang, Naomi Leonard, Adji Bousso Dieng, Christine Allen-Blanchette

From pedestrians to Kuramoto oscillators, interactions between agents govern how a multitude of dynamical systems evolve in space and time. Discovering how these agents relate to each other can improve our understanding of the often complex dynamics that underlie these systems. Recent works learn to categorize relationships between agents based on observations of their physical behavior. These approaches are limited in that the relationship categories are modelled as outcomes of categorical distribution, when in real world systems categories often intermingle and interact. In this work, we introduce a level of abstraction between the observable behavior of agents and the latent categories that determine their behavior. To do this, we learn a mapping from agent behavior to agent preferences for each latent category in a graph neural network. We integrate the physical proximity of agents and their preferences in a nonlinear opinion dynamics model which provides a mechanism to identify mutually exclusive latent categories, predict an agent's evolution in time, and control an agent's physical behavior. We demonstrate the utility of our model for learning interpretable categories, and its efficacy on long-horizon prediction across several benchmarks where we outperform existing methods.

☆ Agent-to-Sim: Learning Interactive Behavior Models from Casual Longitudinal Videos

Gengshan Yang, Andrea Bajcsy, Shunsuke Saito, Angjoo Kanazawa

We present Agent-to-Sim (ATS), a framework for learning interactive behavior models of 3D agents from casual longitudinal video collections. Different from prior works that rely on marker-based tracking and multiview cameras, ATS learns natural behaviors of animal and human agents non-invasively through video observations recorded over a long time-span (e.g., a month) in a single environment. Modeling 3D behavior of an agent requires persistent 3D tracking (e.g., knowing which point corresponds to which) over a long time period. To obtain such data, we develop a coarse-to-fine registration method that tracks the agent and the camera over time through a canonical 3D space, resulting in a complete and persistent spacetime 4D representation. We then train a generative model of agent behaviors using paired data of perception and motion of an agent queried from the 4D reconstruction. ATS enables real-to-sim transfer from video recordings of an agent to an interactive behavior simulator. We demonstrate results on pets (e.g., cat, dog, bunny) and human given monocular RGBD videos captured by a smartphone.

comment: Project page: https://gengshan-y.github.io/agent2sim-www/

☆ CoT-TL: Low-Resource Temporal Knowledge Representation of Planning Instructions Using Chain-of-Thought Reasoning IROS 2024

Kumar Manas, Stefan Zwicklbauer, Adrian Paschke

Autonomous agents often face the challenge of interpreting uncertain natural language instructions for planning tasks. Representing these instructions as Linear Temporal Logic (LTL) enables planners to synthesize actionable plans. We introduce CoT-TL, a data-efficient in-context learning framework for translating natural language specifications into LTL representations. CoT-TL addresses the limitations of large language models, which typically rely on extensive fine-tuning data, by extending chain-of-thought reasoning and semantic roles to align with the requirements of formal logic creation. This approach enhances the transparency and rationale behind LTL generation, fostering user trust. CoT-TL achieves state-of-the-art accuracy across three diverse datasets in low-data scenarios, outperforming existing methods without fine-tuning or intermediate translations. To improve reliability and minimize hallucinations, we incorporate model checking to validate the syntax of the generated LTL output. We further demonstrate CoT-TL's effectiveness through ablation studies and evaluations on unseen LTL structures and formulas in a new dataset. Finally, we validate CoT-TL's practicality by integrating it into a QuadCopter for multi-step drone planning based on natural language instructions.

comment: Accepted for publication in Proceedings of the 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2024), Abu Dhabi 14-18 October 2024

★ LASER: Script Execution by Autonomous Agents for On-demand Traffic Simulation

Hao Gao, Jingyue Wang, Wenyang Fang, Jingwei Xu, Yunpeng Huang, Taolue Chen, Xiaoxing Ma

Autonomous Driving Systems (ADS) require diverse and safety-critical traffic scenarios for effective training and testing, but the existing data generation methods struggle to provide flexibility and scalability. We propose LASER, a novel frame-work that leverage large language models (LLMs) to conduct traffic simulations based on natural language inputs. The framework operates in two stages: it first generates scripts from user-provided descriptions and then executes them using autonomous agents in real time. Validated in the CARLA simulator, LASER successfully generates complex, on-demand driving scenarios, significantly improving ADS training and testing data generation.

☆ Continuum Robot Shape Estimation Using Magnetic Ball Chains

Giovanni Pittiglio, Abdulhamit Donder, Pierre E. Dupont

Shape sensing of medical continuum robots is important both for closed-loop control as well as for enabling the clinician to visualize the robot inside the body. There is a need for inexpensive, but accurate shape sensing technologies. This paper proposes the use of magnetic ball chains as a means of generating shape-specific magnetic fields that can be detected by an external array of Hall effect sensors. Such a ball chain, encased in a flexible polymer sleeve, could be inserted inside the lumen of any continuum robot to provide real-time shape feedback. The sleeve could be removed, as needed, during the procedure to enable use of the entire lumen. To investigate this approach, a shape-sensing model for a steerable catheter tip is derived and an observability and sensitivity analysis are presented. Experiments show maximum estimation errors of 7.1% and mean of 2.9% of the tip position with respect to total length.

☆ ARCADE: Scalable Demonstration Collection and Generation via Augmented Reality for Imitation Learning

Yue Yang, Bryce Ikeda, Gedas Bertasius, Daniel Szafir

Robot Imitation Learning (IL) is a crucial technique in robot learning, where agents learn by mimicking human demonstrations. However, IL encounters scalability challenges stemming from both non-user-friendly demonstration collection methods and the extensive time required to amass a sufficient number of demonstrations for effective training. In response, we introduce the Augmented Reality for Collection and generAtion of DEmonstrations (ARCADE) framework, designed to scale up demonstration collection for robot manipulation tasks. Our framework combines two key capabilities: 1) it leverages AR to make demonstration collection as simple as users performing daily tasks using their hands, and 2) it enables the automatic generation of additional synthetic demonstrations from a single human-derived demonstration, significantly reducing user effort and time. We assess ARCADE's performance on a real Fetch robot across three robotics tasks: 3-Waypoints-Reach, Push, and Pick-And-Place. Using our framework, we were able to rapidly train a policy using vanilla Behavioral Cloning (BC), a classic IL algorithm, which excelled across these three tasks. We also deploy ARCADE on a real household task, Pouring-Water, achieving an 80% success rate.

☆ Analyzing Closed-loop Training Techniques for Realistic Traffic Agent Models in Autonomous Highway Driving Simulations

Matthias Bitzer, Reinis Cimurs, Benjamin Coors, Johannes Goth, Sebastian Ziesche, Philipp Geiger, Maximilian Naumann

Simulation plays a crucial role in the rapid development and safe deployment of autonomous vehicles. Realistic traffic agent models are indispensable for bridging the gap between simulation and the real world. Many existing approaches for imitating human behavior are based on learning from demonstration. However, these approaches are often constrained by focusing on individual training strategies. Therefore, to foster a broader understanding of realistic traffic agent modeling, in this paper, we provide an extensive comparative analysis of different training principles, with a focus on closed-loop methods for highway driving simulation. We experimentally compare (i) open-loop vs. closed-loop multi-agent training, (ii) adversarial vs. deterministic supervised training, (iii) the impact of reinforcement losses, and (iv) the impact of training alongside log-replayed agents to identify suitable training techniques for realistic agent modeling. Furthermore, we identify promising combinations of different closed-loop training methods.

comment: 15 pages, 6 figures, 4 tables

★ Learning Quadrotor Control From Visual Features Using Differentiable Simulation

Johannes Heeg, Yunlong Song, Davide Scaramuzza

The sample inefficiency of reinforcement learning (RL) remains a significant challenge in robotics. RL requires large-scale simulation and, still, can cause long training times, slowing down research and innovation. This issue is particularly pronounced in vision-based control tasks where reliable state estimates are not accessible. Differentiable simulation offers an alternative by enabling gradient back-propagation through the dynamics model, providing low-variance analytical policy gradients and, hence, higher sample efficiency. However, its usage for real-world robotic tasks has yet been limited. This work demonstrates the great potential of differentiable simulation for learning quadrotor control. We show that training in differentiable simulation significantly outperforms model-free RL in terms of both sample efficiency and training time, allowing a policy to learn to recover a quadrotor in seconds when providing vehicle state and in minutes when relying solely on visual features. The key to our success is two-fold. First, the use of a simple surrogate model for gradient computation greatly accelerates training without sacrificing control performance. Second, combining state representation learning with policy learning enhances convergence speed in tasks where only visual features are observable. These findings highlight the potential of differentiable simulation for real-world robotics and offer a compelling alternative to conventional RL approaches.

comment: Under Submission

☆ Diffusion Transformer Policy

Zhi Hou, Tianyi Zhang, Yuwen Xiong, Hengjun Pu, Chengyang Zhao, Ronglei Tong, Yu Qiao, Jifeng Dai, Yuntao Chen

Recent large visual-language action models pretrained on diverse robot datasets have demonstrated the potential for generalizing to new environments with a few in-domain data. However, those approaches usually predict discretized or continuous actions by a small action head, which limits the ability in handling diverse action spaces. In contrast, we model the continuous action with a large multi-modal diffusion transformer, dubbed as Diffusion Transformer Policy, in which we directly denoise action chunks by a large transformer model rather than a small action head. By leveraging the scaling capability of transformers, the proposed approach can effectively model continuous end-effector actions across large diverse robot datasets, and achieve better generalization performance. Extensive experiments demonstrate Diffusion Transformer Policy pretrained on diverse robot data can generalize to different embodiments, including simulation environments like Maniskill2 and Calvin, as well as the real-world Franka arm. Specifically, without bells and whistles, the proposed approach achieves state-of-the-art performance with only a single third-view camera stream in the Calvin novel task setting (ABC->D), improving the average number of tasks completed in a row of 5 to 3.6, and the pretraining stage significantly facilitates the success sequence length on the Calvin by over 1.2. The code will be publicly available.

comment: Preprint

☆ Neural Predictor for Flight Control with Payload

Ao Jin, Chenhao Li, Qinyi Wang, Ya Liu, Panfeng Huang, Fan Zhang

Aerial robotics for transporting suspended payloads as the form of freely-floating manipulator are growing great interest in recent years. However, the prior information of the payload, such as the mass, is always hard to obtain accurately in practice. The force/torque caused by payload and residual dynamics will introduce unmodeled perturbations to the system, which negatively affects the closed-loop performance. Different from estimation-like methods, this paper proposes Neural Predictor, a learning-based approach to model force/torque caused by payload and residual dynamics as a dynamical system. It results a hybrid model including both the first-principles dynamics and the learned dynamics. This hybrid model is then integrated into a MPC framework to improve closed-loop performance. Effectiveness of proposed framework is verified extensively in both numerical simulations and real-world flight experiments. The results indicate that our approach can capture force/torque caused by payload and residual dynamics accurately, respond quickly to the changes of them and improve the closed-loop performance significantly. In particular, Neural Predictor outperforms a state-of-the-art learning-based estimator and has reduced the force and torque estimation errors by up to 66.15% and 33.33% while using less samples.

comment: 8 pages

☆ Fully distributed and resilient source seeking for robot swarms

Jesús Bautista, Antonio Acuaviva, José Hinojosa, Weijia Yao, Juan Jiménez, Héctor García de Marina

We propose a self-contained, resilient and fully distributed solution for locating the maximum of an unknown 3D scalar field using a swarm of robots that travel at constant speeds. Unlike conventional reactive methods relying on gradient information, our methodology enables the swarm to determine an ascending direction so that it approaches the source with arbitrary precision. Our source-seeking solution consists of three algorithms. The first two algorithms run sequentially and distributively at a high frequency providing barycentric coordinates and the ascending direction respectively to the individual robots. The third algorithm is the individual control law for a robot to track the estimated ascending direction. We show that the two algorithms with higher frequency have an exponential convergence to their eventual values since they are based on the standard consensus protocol for first-order dynamical systems; their high frequency depends on how fast the robots travel through the scalar field. The robots are not constrained to any particular geometric formation, and we study both discrete and continuous distributions of robots within swarm shapes. The shape analysis reveals the resiliency of our approach as expected in robot swarms, i.e., by amassing robots we ensure the source-seeking functionality in the event of missing or misplaced individuals or even if the robot network splits into two or more disconnected subnetworks. In addition, we also enhance the robustness of the algorithm by presenting conditions for \emph{optimal} swarm shapes, in the sense that the ascending directions can be closely parallel to the field's gradient. We exploit such an analysis so that the swarm can adapt to unknown environments by morphing its shape and maneuvering while still following an ascending direction.

comment: 15 pages, submitted version to T-RO. This version does not contain the field experiments. arXiv admin note: text overlap with arXiv:2309.02937

☆ Bench4Merge: A Comprehensive Benchmark for Merging in Realistic Dense Traffic with Micro-Interactive Vehicles

Zhengming Wang, Junli Wang, Pengfei Li, Zhaohan Li, Peng Li, Yilun Chen

While the capabilities of autonomous driving have advanced rapidly, merging into dense traffic remains a significant challenge, many motion planning methods for this scenario have been proposed but it is hard to evaluate them. Most existing closed-loop simulators rely on rule-based controls for other vehicles, which results in a lack of diversity and randomness, thus failing to accurately assess the motion planning capabilities in highly interactive scenarios. Moreover, traditional evaluation metrics are insufficient for comprehensively evaluating the performance of merging in dense traffic. In response, we proposed a closed-loop evaluation benchmark for assessing motion planning capabilities in merging scenarios. Our approach involves other vehicles trained in large scale datasets with micro-behavioral characteristics that significantly enhance the complexity and diversity. Additionally, we have restructured the evaluation mechanism by leveraging large language models to assess each autonomous vehicle merging onto the main road. Extensive experiments have demonstrated the advanced nature of this evaluation benchmark. Through this benchmark, we have obtained an evaluation of existing methods and identified common issues. The environment and vehicle motion planning models we have designed can be accessed at https://anonymous.4open.science/r/Bench4Merge-EB5D

comment: 6 pages, 7 figures, IEEE international conference on robotics and automation

☆ Distributed Learning for UAV Swarms

Chen Hu, Hanchi Ren, Jingjing Deng, Xianghua Xie

Unmanned Aerial Vehicle (UAV) swarms are increasingly deployed in dynamic, data-rich environments for applications such as environmental monitoring and surveillance. These scenarios demand efficient data processing while maintaining privacy and security, making Federated Learning (FL) a promising solution. FL allows UAVs to collaboratively train global models without sharing raw data, but challenges arise due to the non-Independent and Identically Distributed (non-IID) nature of the data collected by UAVs. In this study, we show an integration of the state-of-the-art FL methods to UAV Swarm application and invetigate the performance of multiple aggregation methods (namely FedAvg, FedProx, FedOpt, and MOON) with a particular focus on tackling non-IID on a variety of datasets, specifically MNIST for baseline performance, CIFAR10 for natural object classification, EuroSAT for environment monitoring, and CelebA for surveillance. These algorithms were selected to cover improved techniques on both client-side updates and global aggregation. Results show that while all algorithms perform comparably on IID data, their performance deteriorates significantly under non-IID conditions. FedProx demonstrated the most stable overall performance, emphasising the importance of regularising local updates in non-IID environments to mitigate drastic deviations in local models.

☆ Triplane Grasping: Efficient 6-DoF Grasping with Single RGB Images

Yiming Li, Hanchi Ren, Jingjing Deng, Xianghua Xie

Reliable object grasping is one of the fundamental tasks in robotics. However, determining grasping pose based on single-image input has long been a challenge due to limited visual information and the complexity of real-world objects. In this paper, we propose Triplane Grasping, a fast grasping decision-making method that relies solely on a single RGB-only image as input. Triplane Grasping creates a hybrid Triplane-Gaussian 3D representation through a point decoder and a triplane decoder, which produce an efficient and high-quality reconstruction of the object to be grasped to meet real-time grasping requirements. We propose to use an end-to-end network to generate 6-DoF parallel-jaw grasp distributions directly from 3D points in the point cloud as potential grasp contacts and anchor the grasp pose in the observed data. Experiments demonstrate that our method achieves rapid modeling and grasping pose decision-making for daily objects, and exhibits a high grasping success rate in zero-shot scenarios.

☆ Safety-critical Control with Control Barrier Functions: A Hierarchical Optimization Framework

Junjun Xie, Liang Hu, Jiahu Qin, Jun Yang, Huijun Gao

The control barrier function (CBF) has become a fundamental tool in safety-critical systems design since its invention. Typically, the quadratic optimization framework is employed to accommodate CBFs, control Lyapunov functions (CLFs), other constraints and nominal control design. However, the constrained optimization framework involves hyper-parameters to tradeoff different objectives and constraints, which, if not well-tuned beforehand, impact system performance and even lead to infeasibility. In this paper, we propose a hierarchical optimization framework that decomposes the multi-objective optimization problem into nested optimization sub-problems in a safety-first approach. The new framework addresses potential infeasibility on the premise of ensuring safety and performance as much as possible and applies easily in multi-certificate cases. With vivid visualization aids, we systematically analyze the advantages of our proposed method over existing QP-based ones in terms of safety, feasibility and convergence rates. Moreover, two numerical examples are provided that verify our analysis and show the superiority of our proposed method.

☆ Robust Loop Closure by Textual Cues in Challenging Environments

Tongxing Jin, Thien-Minh Nguyen, Xinhang Xu, Yizhuo Yang, Shenghai Yuan, Jianping Li, Lihua Xie

Loop closure is an important task in robot navigation. However, existing methods mostly rely on some implicit or heuristic features of the environment, which can still fail to work in common environments such as corridors, tunnels, and warehouses. Indeed, navigating in such featureless, degenerative, and repetitive (FDR) environments would also pose a significant challenge even for humans, but explicit text cues in the surroundings often provide the best assistance. This inspires us to propose a multi-modal loop closure method based on explicit human-readable textual cues in FDR environments. Specifically, our approach first extracts scene text entities based on Optical Character Recognition (OCR), then creates a local map of text cues based on accurate LiDAR odometry and finally identifies loop closure events by a graph-theoretic scheme. Experiment results demonstrate that this approach has superior performance over existing methods that rely solely on visual and LiDAR sensors. To benefit the community, we release the source code and datasets at \url{https://github.com/TongxingJin/TXTLCD}.

☆ Task-oriented Robotic Manipulation with Vision Language Models

Nurhan Bulus Guran, Hanchi Ren, Jingjing Deng, Xianghua Xie

Vision-Language Models (VLMs) play a crucial role in robotic manipulation by enabling robots to understand and interpret the visual properties of objects and their surroundings, allowing them to perform manipulation based on this multimodal understanding. However, understanding object attributes and spatial relationships is a non-trivial task but is critical in robotic manipulation tasks. In this work, we present a new dataset focused on spatial relationships and attribute assignment and a novel method to utilize VLMs to perform object manipulation with task-oriented, high-level input. In this dataset, the spatial relationships between objects are manually described as captions. Additionally, each object is labeled with multiple attributes, such as fragility, mass, material, and transparency, derived from a fine-tuned vision language model. The embedded object information from captions are automatically extracted and transformed into a data structure (in this case, tree, for demonstration purposes) that captures the spatial relationships among the objects within each image. The tree structures, along with the object attributes, are then fed into a language model to transform into a new tree structure that determines how these objects should be organized in order to accomplish a specific (high-level) task. We demonstrate that our method not only improves the comprehension of spatial relationships among objects in the visual environment but also enables robots to interact with these objects more effectively. As a result, this approach significantly enhances spatial reasoning in robotic manipulation tasks. To our knowledge, this is the first method of its kind in the literature, offering a novel solution that allows robots to more efficiently organize and utilize objects in their surroundings.

☆ Long-distance Geomagnetic Navigation in GNSS-denied Environments with Deep Reinforcement Learning

Wenqi Bai, Xiaohui Zhang, Shiliang Zhang, Songnan Yang, Yushuai Li, Tingwen Huang

Geomagnetic navigation has drawn increasing attention with its capacity in navigating through complex environments and its independence from external navigation services like global navigation satellite systems (GNSS). Existing studies on geomagnetic navigation, i.e., matching navigation and bionic navigation, rely on pre-stored map or extensive searches, leading to limited applicability or reduced navigation efficiency in unexplored areas. To address the issues with geomagnetic navigation in areas where GNSS is unavailable, this paper develops a deep reinforcement learning (DRL)-based mechanism, especially for long-distance geomagnetic navigation. The designed mechanism trains an agent to learn and gain the magnetoreception capacity for geomagnetic navigation, rather than using any pre-stored map or extensive and expensive searching approaches. Particularly, we integrate the geomagnetic gradient-based parallel approach into geomagnetic navigation. This integration mitigates the over-exploration of the learning agent by adjusting the geomagnetic gradient, such that the obtained gradient is aligned towards the destination. We explore the effectiveness of the proposed approach via detailed numerical simulations, where we implement twin delayed deep deterministic policy gradient (TD3) in realizing the proposed approach. The results demonstrate that our approach outperforms existing metaheuristic and bionic navigation methods in long-distance missions under diverse navigation conditions.

☆ Assisted Physical Interaction: Autonomous Aerial Robots with Neural Network Detection, Navigation, and Safety Layers

Andrea Berra, Viswa Narayanan Sankaranarayanan, Achilleas Santi Seisa, Julien Mellet, Udayanga G. W. K. N. Gamage, Sumeet Gajanan Satpute, Fabio Ruggiero, Vincenzo Lippiello, Silvia Tolu, Matteo Fumagalli, George Nikolakopoulos, Miguel Ángel Trujillo Soto, Guillermo Heredia

The paper introduces a novel framework for safe and autonomous aerial physical interaction in industrial settings. It comprises two main components: a neural network-based target detection system enhanced with edge computing for reduced onboard computational load, and a control barrier function (CBF)-based controller for safe and precise maneuvering. The target detection system is trained on a dataset under challenging visual conditions and evaluated for accuracy across various unseen data with changing lighting conditions. Depth features are utilized for target pose estimation, with the entire detection framework offloaded into low-latency edge computing. The CBF-based controller enables the UAV to converge safely to the target for precise contact. Simulated evaluations of both the controller and target detection are presented, alongside an analysis of real-world detection performance.

comment: 8 pages,14 figures, ICUAS 2024

☆ Flying through Moving Gates without Full State Estimation

Ralf Römer, Tim Emmert, Angela P. Schoellig

Autonomous drone racing requires powerful perception, planning, and control and has become a benchmark and test field for autonomous, agile flight. Existing work usually assumes static race tracks with known maps, which enables offline planning of time-optimal trajectories, performing localization to the gates to reduce the drift in visual-inertial odometry (VIO) for state estimation or training learning-based methods for the particular race track and operating environment. In contrast, many real-world tasks like disaster response or delivery need to be performed in unknown and dynamic environments. To close this gap and make drone racing more robust against unseen environments and moving gates, we propose a control algorithm that does not require a race track map or VIO and uses only monocular measurements of the line of sight (LOS) to the gates. For this purpose, we adopt the law of proportional navigation (PN) to accurately fly through the gates despite gate motions or wind. We formulate the PN-informed vision-based control problem for drone racing as a constrained optimization problem and derive a closed-form optimal solution. We demonstrate through extensive simulations and real-world experiments that our method can navigate through moving gates at high speeds while being robust to different gate movements, model errors, wind, and delays.

comment: 7 pages, 6 figures

☆ Design of a Flexible Robot Arm for Safe Aerial Physical Interaction

Julien Mellet, Andrea Berra, Achilleas Santi Seisa, Viswa Sankaranarayanan, Udayanga G. W. K. N. Gamage, Miguel Angel Trujillo Soto, Guillermo Heredia, George Nikolakopoulos, Vincenzo Lippiello, Fabio Ruggiero

This paper introduces a novel compliant mechanism combining lightweight and energy dissipation for aerial physical interaction. Weighting 400~g at take-off, the mechanism is actuated in the forward body direction, enabling precise position control for force interaction and various other aerial manipulation tasks. The robotic arm, structured as a closed-loop kinematic chain, employs two deported servomotors. Each joint is actuated with a single tendon for active motion control in compression of the arm at the end-effector. Its elasto-mechanical design reduces weight and provides flexibility, allowing passive-compliant interactions without impacting the motors' integrity. Notably, the arm's damping can be adjusted based on the proposed inner frictional bulges. Experimental applications showcase the aerial system performance in both free-flight and physical interaction. The presented work may open safer applications for \ac{MAV} in real environments subject to perturbations during interaction.

comment: 6 pages, 7 figures, ROBOSOFT 2024

☆ WildOcc: A Benchmark for Off-Road 3D Semantic Occupancy Prediction

Heng Zhai, Jilin Mei, Chen Min, Liang Chen, Fangzhou Zhao, Yu Hu

3D semantic occupancy prediction is an essential part of autonomous driving, focusing on capturing the geometric details of scenes. Off-road environments are rich in geometric information, therefore it is suitable for 3D semantic occupancy prediction tasks to reconstruct such scenes. However, most of researches concentrate on on-road environments, and few methods are designed for off-road 3D semantic occupancy prediction due to the lack of relevant datasets and benchmarks. In response to this gap, we introduce WildOcc, to our knowledge, the first benchmark to provide dense occupancy annotations for off-road 3D semantic occupancy prediction tasks. A ground truth generation pipeline is proposed in this paper, which employs a coarse-to-fine reconstruction to achieve a more realistic result. Moreover, we introduce a multi-modal 3D semantic occupancy prediction framework, which fuses spatio-temporal information from multi-frame images and point clouds at voxel level. In addition, a cross-modality distillation function is introduced, which transfers geometric knowledge from point clouds to image features.

★ Generalizing Motion Planners with Mixture of Experts for Autonomous Driving

Qiao Sun, Huimin Wang, Jiahao Zhan, Fan Nie, Xin Wen, Leimeng Xu, Kun Zhan, Peng Jia, Xianpeng Lang, Hang Zhao

Large real-world driving datasets have sparked significant research into various aspects of data-driven motion planners for autonomous driving. These include data augmentation, model architecture, reward design, training strategies, and planner pipelines. These planners promise better generalizations on complicated and few-shot cases than previous methods. However, experiment results show that many of these approaches produce limited generalization abilities in planning performance due to overly complex designs or training paradigms. In this paper, we review and benchmark previous methods focusing on generalizations. The experimental results indicate that as models are appropriately scaled, many design elements become redundant. We introduce StateTransformer-2 (STR2), a scalable, decoder-only motion planner that uses a Vision Transformer (ViT) encoder and a mixture-of-experts (MoE) causal Transformer architecture. The MoE backbone addresses modality collapse and reward balancing by expert routing during training. Extensive experiments on the NuPlan dataset show that our method generalizes better than previous approaches across different test sets and closed-loop simulations. Furthermore, we assess its scalability on billions of real-world urban driving scenarios, demonstrating consistent accuracy improvements as both data and model size grow.

comment: 7 pages, 3 figures

☆ MSGField: A Unified Scene Representation Integrating Motion, Semantics, and Geometry for Robotic Manipulation

Yu Sheng, Runfeng Lin, Lidian Wang, Quecheng Qiu, YanYong Zhang, Yu Zhang, Bei Hua, Jianmin Ji

Combining accurate geometry with rich semantics has been proven to be highly effective for language-guided robotic manipulation. Existing methods for dynamic scenes either fail to update in real-time or rely on additional depth sensors for simple scene editing, limiting their applicability in real-world. In this paper, we introduce MSGField, a representation that uses a collection of 2D Gaussians for high-quality reconstruction, further enhanced with attributes to encode semantic and motion information. Specially, we represent the motion field compactly by decomposing each primitive's motion into a combination of a limited set of motion bases. Leveraging the differentiable real-time rendering of Gaussian splatting, we can quickly optimize object motion, even for complex non-rigid motions, with image supervision from only two camera views. Additionally, we designed a pipeline that utilizes object priors to efficiently obtain well-defined semantics. In our challenging dataset, which includes flexible and extremely small objects, our method achieve a success rate of 79.2% in static and 63.3% in dynamic environments for language-guided manipulation. For specified object grasping, we achieve a success rate of 90%, on par with point cloud-based methods. Code and dataset will be released at:https://shengyu724.github.io/MSGField.github.io.

☆ Efficient Non-Myopic Layered Bayesian Optimization For Large-Scale Bathymetric Informative Path Planning ICRA

Alexander Kiessling, Ignacio Torroba, Chelsea Rose Sidrane, Ivan Stenius, Jana Tumova, John Folkesson

Informative path planning (IPP) applied to bathymetric mapping allows AUVs to focus on feature-rich areas to quickly reduce uncertainty and increase mapping efficiency. Existing methods based on Bayesian optimization (BO) over Gaussian Process (GP) maps work well on small scenarios but they are short-sighted and computationally heavy when mapping larger areas, hindering deployment in real applications. To overcome this, we present a 2-layered BO IPP method that performs non-myopic, real-time planning in a tree search fashion over large Stochastic Variational GP maps, while respecting the AUV motion constraints and accounting for localization uncertainty. Our framework outperforms the standard industrial lawn-mowing pattern and a myopic baseline in a set of hardware in the loop (HIL) experiments in an embedded platform over real bathymetry.

comment: 6 pages + 1 page of references, 4 figures, submitted to International Conference on Robotics and Automation (ICRA)

☆ Hierarchical Search-Based Cooperative Motion Planning

Yuchen Wu, Yifan Yang, Gang Xu, Junjie Cao, Yansong Chen, Licheng Wen, Yong Liu

Cooperative path planning, a crucial aspect of multi-agent systems research, serves a variety of sectors, including military, agriculture, and industry. Many existing algorithms, however, come with certain limitations, such as simplified kinematic models and inadequate support for multiple group scenarios. Focusing on the planning problem associated with a nonholonomic Ackermann model for Unmanned Ground Vehicles (UGV), we propose a leaderless, hierarchical Search-Based Cooperative Motion Planning (SCMP) method. The high-level utilizes a binary conflict search tree to minimize runtime, while the low-level fabricates kinematically feasible, collision-free paths that are shape-constrained. Our algorithm can adapt to scenarios featuring multiple groups with different shapes, outlier agents, and elaborate obstacles. We conduct algorithm comparisons, performance testing, simulation, and real-world testing, verifying the effectiveness and applicability of our algorithm. The implementation of our method will be open-sourced at https://github.com/WYCUniverStar/SCMP.

☆ PALMS: Plane-based Accessible Indoor Localization Using Mobile Smartphones

Yunqian Cheng, Roberto Manduchi

In this paper, we present PALMS, an innovative indoor global localization and relocalization system for mobile smartphones that utilizes publicly available floor plans. Unlike most vision-based methods that require constant visual input, our system adopts a dynamic form of localization that considers a single instantaneous observation and odometry data. The core contribution of this work is the introduction of a particle filter initialization method that leverages the Certainly Empty Space (CES) constraint along with principal orientation matching. This approach creates a spatial probability distribution of the device's location, significantly improving localization accuracy and reducing particle filter convergence time. Our experimental evaluations demonstrate that PALMS outperforms traditional methods with uniformly initialized particle filters, providing a more efficient and accessible approach to indoor wayfinding. By eliminating the need for prior environmental fingerprinting, PALMS provides a scalable and practical approach to indoor navigation.

comment: 7 pages, 3 figures, accepted to the 14th International Conference on Indoor Positioning and Indoor Navigation (IPIN) 2024, Best Presentation Award

☆ RANSAC Back to SOTA: A Two-stage Consensus Filtering for Real-time 3D Registration

Pengcheng Shi, Shaocheng Yan, Yilin Xiao, Xinyi Liu, Yongjun Zhang, Jiayuan Li

Correspondence-based point cloud registration (PCR) plays a key role in robotics and computer vision. However, challenges like sensor noises, object occlusions, and descriptor limitations inevitably result in numerous outliers. RANSAC family is the most popular outlier removal solution. However, the requisite iterations escalate exponentially with the outlier ratio, rendering it far inferior to existing methods (SC2PCR [1], MAC [2], etc.) in terms of accuracy or speed. Thus, we propose a two-stage consensus filtering (TCF) that elevates RANSAC to state-of-the-art (SOTA) speed and accuracy. Firstly, one-point RANSAC obtains a consensus set based on length consistency. Subsequently, two-point RANSAC refines the set via angle consistency. Then, three-point RANSAC computes a coarse pose and removes outliers based on transformed correspondence's distances. Drawing on optimizations from one-point and two-point RANSAC, three-point RANSAC requires only a few iterations. Eventually, an iterative reweighted least squares (IRLS) is applied to yield the optimal pose. Experiments on the large-scale KITTI and ETH datasets demonstrate our method achieves up to three-orders-of-magnitude speedup compared to MAC while maintaining registration accuracy and recall. Our code is available at https://github.com/ShiPC-AI/TCF.

comment: 8 pages, 8 figures

☆ Reinforced Imitative Trajectory Planning for Urban Automated Driving

Di Zeng, Ling Zheng, Xiantong Yang, Yinong Li

Reinforcement learning (RL) faces challenges in trajectory planning for urban automated driving due to the poor convergence of RL and the difficulty in designing reward functions. The convergence problem is alleviated by combining RL with supervised learning. However, most existing approaches only reason one step ahead and lack the capability to plan for multiple future steps. Besides, although inverse reinforcement learning holds promise for solving the reward function design issue, existing methods for automated driving impose a linear structure assumption on reward functions, making them difficult to apply to urban automated driving. In light of these challenges, this paper proposes a novel RL-based trajectory planning method that integrates RL with imitation learning to enable multi-step planning. Furthermore, a transformer-based Bayesian reward function is developed, providing effective reward signals for RL in urban scenarios. Moreover, a hybrid-driven trajectory planning framework is proposed to enhance safety and interpretability. The proposed methods were validated on the large-scale real-world urban automated driving nuPlan dataset. The results demonstrated the significant superiority of the proposed methods over the baselines in terms of the closed-loop metrics. The code is available at https://github.com/Zigned/nuplan_zigned.

comment: 19 pages, 9 figures

☆ Patrol Security Game: Defending Against Adversary with Freedom in Attack Timing, Location, and Duration

Hao-Tsung Yang, Ting-Kai Weng, Ting-Yu Chang, Kin Sum Liu, Shan Lin, Jie Gao, Shih-Yu Tsai

We explored the Patrol Security Game (PSG), a robotic patrolling problem modeled as an extensive-form Stackelberg game, where the attacker determines the timing, location, and duration of their attack. Our objective is to devise a patrolling schedule with an infinite time horizon that minimizes the attacker's payoff. We demonstrated that PSG can be transformed into a combinatorial minimax problem with a closed-form objective function. By constraining the defender's strategy to a time-homogeneous first-order Markov chain (i.e., the patroller's next move depends solely on their current location), we proved that the optimal solution in cases of zero penalty involves either minimizing the expected hitting time or return time, depending on the attacker model, and that these solutions can be computed efficiently. Additionally, we observed that increasing the randomness in the patrol schedule reduces the attacker's expected payoff in high-penalty cases. However, the minimax problem becomes non-convex in other scenarios. To address this, we formulated a bi-criteria optimization problem incorporating two objectives: expected maximum reward and entropy. We proposed three graph-based algorithms and one deep reinforcement learning model, designed to efficiently balance the trade-off between these two objectives. Notably, the third algorithm can identify the optimal deterministic patrol schedule, though its runtime grows exponentially with the number of patrol spots. Experimental results validate the effectiveness and scalability of our solutions, demonstrating that our approaches outperform state-of-the-art baselines on both synthetic and real-world crime datasets.

comment: Under review of TCPS

☆ Development of Minimal Biorobotic Stealth Distance and Its Application in the Design of Direct-Drive Dragonfly-Inspired Aircraft

Zhang Minghao, Song Bifeng, Yang Xiaojun, Wang Liang, Lang Xinyua

This paper introduces the Minimal Biorobotic Stealth Distance (MBSD), a novel quantitative metric to evaluate the bionic resemblance of biorobotic aircraft. Current technological limitations prevent dragonfly-inspired aircrafts from achieving optimal performance at biological scales. To address these challenges, we use the DDD-1 dragonfly-inspired aircraft, a hover-capable direct-drive aircraft, to explore the impact of the MBSD on aircraft design. Key contributions of this research include: (1) the establishment of the MBSD as a quantifiable and operable evaluation metric that influences aircraft design, integrating seamlessly with the overall design process and providing a new dimension for optimizing bionic aircraft, balancing mechanical attributes and bionic characteristics; (2) the creation and analysis of a typical aircraft in four directions: essential characteristics of the MBSD, its coupling relationship with existing performance metrics (Longest Hover Duration and Maximum Instantaneous Forward Flight Speed), multi-objective optimization, and application in a typical mission scenario; (3) the construction and validation of a full-system model for the direct-drive dragonfly-inspired aircraft, demonstrating the design model's effectiveness against existing aircraft data. Detailed calculations of the MBSD consider appearance similarity, dynamic similarity, and environmental similarity.

comment: 61 pages, 32 figures

☆ A Plug-and-Play Fully On-the-Job Real-Time Reinforcement Learning Algorithm for a Direct-Drive Tandem-Wing Experiment Platforms Under Multiple Random Operating Conditions

Zhang Minghao, Song Bifeng, Yang Xiaojun, Wang Liang

The nonlinear and unstable aerodynamic interference generated by the tandem wings of such biomimetic systems poses substantial challenges for motion control, especially under multiple random operating conditions. To address these challenges, the Concerto Reinforcement Learning Extension (CRL2E) algorithm has been developed. This plug-and-play, fully on-the-job, real-time reinforcement learning algorithm incorporates a novel Physics-Inspired Rule-Based Policy Composer Strategy with a Perturbation Module alongside a lightweight network optimized for real-time control. To validate the performance and the rationality of the module design, experiments were conducted under six challenging operating conditions, comparing seven different algorithms. The results demonstrate that the CRL2E algorithm achieves safe and stable training within the first 500 steps, improving tracking accuracy by 14 to 66 times compared to the Soft Actor-Critic, Proximal Policy Optimization, and Twin Delayed Deep Deterministic Policy Gradient algorithms. Additionally, CRL2E significantly enhances performance under various random operating conditions, with improvements in tracking accuracy ranging from 8.3% to 60.4% compared to the Concerto Reinforcement Learning (CRL) algorithm. The convergence speed of CRL2E is 36.11% to 57.64% faster than the CRL algorithm with only the Composer Perturbation and 43.52% to 65.85% faster than the CRL algorithm when both the Composer Perturbation and Time-Interleaved Capability Perturbation are introduced, especially in conditions where the standard CRL struggles to converge. Hardware tests indicate that the optimized lightweight network structure excels in weight loading and average inference time, meeting real-time control requirements.

comment: 63 pages, 32 figures

☆ A Dual Process VLA: Efficient Robotic Manipulation Leveraging VLM

ByungOk Han, Jaehong Kim, Jinhyeok Jang

Vision-Language-Action (VLA) models are receiving increasing attention for their ability to enable robots to perform complex tasks by integrating visual context with linguistic commands. However, achieving efficient real-time performance remains challenging due to the high computational demands of existing models. To overcome this, we propose Dual Process VLA (DP-VLA), a hierarchical framework inspired by dual-process theory. DP-VLA utilizes a Large System 2 Model (L-Sys2) for complex reasoning and decision-making, while a Small System 1 Model (S-Sys1) handles real-time motor control and sensory processing. By leveraging Vision-Language Models (VLMs), the L-Sys2 operates at low frequencies, reducing computational overhead, while the S-Sys1 ensures fast and accurate task execution. Experimental results on the RoboCasa dataset demonstrate that DP-VLA achieves faster inference and higher task success rates, providing a scalable solution for advanced robotic applications.

comment: 10 page

♻ ☆ Knowledge Transfer from Simple to Complex: A Safe and Efficient Reinforcement Learning Framework for Autonomous Driving Decision-Making

Rongliang Zhou, Jiakun Huang, Mingjun Li, Hepeng Li, Haotian Cao, Xiaolin Song

A safe and efficient decision-making system is crucial for autonomous vehicles. However, the complexity and variability of driving environments limit the effectiveness of many rule-based and machine learning-based decision-making approaches. Reinforcement Learning in autonomous driving offers a promising solution to these challenges. Nevertheless, concerns regarding safety and efficiency during training remain major obstacles to its widespread application. To address these concerns, we propose a novel RL framework named Simple to Complex Collaborative Decision. First, we rapidly train the teacher model using the Proximal Policy Optimization algorithm in a lightweight simulation environment. In the more intricate simulation environment, the teacher model intervenes when the student agent exhibits suboptimal behavior by assessing the value of actions to avert dangerous situations. We also introduce an innovative RL algorithm called Adaptive Clipping PPO, which is trained using a combination of samples generated by both teacher and student policies, and employs dynamic clipping strategies based on sample importance. Additionally, we employ the KL divergence as a constraint on policy optimization, transforming it into an unconstrained problem to accelerate the student's learning of the teacher's policy. Finally, a gradual weaning strategy is employed to ensure that, over time, the student agent learns to explore independently. Simulation experiments in highway lane-change scenarios demonstrate that the S2CD framework enhances learning efficiency, reduces training costs, and significantly improves safety during training when compared with state-of-the-art baseline algorithms. This approach also ensures effective knowledge transfer between teacher and student models, and even when the teacher model is suboptimal.

♻ ☆ Bidirectional Decoding: Improving Action Chunking via Closed-Loop Resampling

Yuejiang Liu, Jubayer Ibn Hamid, Annie Xie, Yoonho Lee, Maximilian Du, Chelsea Finn

Predicting and executing a sequence of actions without intermediate replanning, known as action chunking, is increasingly used in robot learning from human demonstrations. Yet, its reported effects on the learned policy are inconsistent: some studies find it crucial for achieving strong results, while others observe decreased performance. In this paper, we first dissect how action chunking impacts the divergence between a learner and a demonstrator. We find that action chunking allows the learner to better capture the temporal dependencies in demonstrations but at the cost of reduced reactivity in stochastic environments. To address this tradeoff, we propose Bidirectional Decoding (BID), a test-time inference algorithm that bridges action chunking with closed-loop operations. BID samples multiple predictions at each time step and searches for the optimal one based on two criteria: (i) backward coherence, which favors samples that align with previous decisions; (ii) forward contrast, which seeks samples of high likelihood for future plans. By coupling decisions within and across action chunks, BID promotes consistency over time while maintaining reactivity to unexpected changes. Experimental results show that BID boosts the performance of two state-of-the-art generative policies across seven simulation benchmarks and two real-world tasks. Code and videos are available at https://bid-robot.github.io.

comment: Project website: https://bid-robot.github.io/

♻ ☆ Data Efficient Behavior Cloning for Fine Manipulation via Continuity-based Corrective Labels IROS 2024

Abhay Deshpande, Liyiming Ke, Quinn Pfeifer, Abhishek Gupta, Siddhartha S. Srinivasa

We consider imitation learning with access only to expert demonstrations, whose real-world application is often limited by covariate shift due to compounding errors during execution. We investigate the effectiveness of the Continuity-based Corrective Labels for Imitation Learning (CCIL) framework in mitigating this issue for real-world fine manipulation tasks. CCIL generates corrective labels by learning a locally continuous dynamics model from demonstrations to guide the agent back toward expert states. Through extensive experiments on peg insertion and fine grasping, we provide the first empirical validation that CCIL can significantly improve imitation learning performance despite discontinuities present in contact-rich manipulation. We find that: (1) real-world manipulation exhibits sufficient local smoothness to apply CCIL, (2) generated corrective labels are most beneficial in low-data regimes, and (3) label filtering based on estimated dynamics model error enables performance gains. To effectively apply CCIL to robotic domains, we offer a practical instantiation of the framework and insights into design choices and hyperparameter selection. Our work demonstrates CCIL's practicality for alleviating compounding errors in imitation learning on physical robots.

comment: Presented at IROS 2024

♻ ☆ A Lyapunov-Based Switching Scheme for Selecting the Stable Closed-Loop Fixed Attitude-Error Quaternion During Flight

Francisco M. F. R. Goncalves, Ryan M. Bena, Konstantin I. Matveev, Nestor O. Perez-Arancibia

We present a switching scheme, which uses both the attitude-error quaternion (AEQ) and the angular-velocity error, for controlling the rotational degrees of freedom of an uncrewed aerial vehicle (UAV) during flight. In this approach, the proposed controller continually selects the stable closed-loop (CL) equilibrium AEQ corresponding to the smallest cost between those computed with two energy-based Lyapunov functions. To analyze and enforce the stability of the CL switching dynamics, we use basic nonlinear theory. This research problem is relevant because the selection of the stable CL equilibrium AEQ directly determines the power and energy requirements of the controlled UAV during flight. To test and demonstrate the implementation, suitability, functionality, and performance of the proposed approach, we present experimental results obtained using a 31-gram quadrotor, which was controlled to execute high-speed yaw maneuvers in flight. These flight tests show that the proposed switching controller can respectively reduce the control effort and rotational power by as much as 49.75 % and 28.14 %, on average, compared to those corresponding to an often-used benchmark controller.

comment: 8 pages, 5 figures, 2024 7th Iberian Robotics Conference (ROBOT)

♻ ☆ Human-Agent Joint Learning for Efficient Robot Manipulation Skill Acquisition

Shengcheng Luo, Quanquan Peng, Jun Lv, Kaiwen Hong, Katherine Rose Driggs-Campbell, Cewu Lu, Yong-Lu Li

Employing a teleoperation system for gathering demonstrations offers the potential for more efficient learning of robot manipulation. However, teleoperating a robot arm equipped with a dexterous hand or gripper, via a teleoperation system presents inherent challenges due to the task's high dimensionality, complexity of motion, and differences between physiological structures. In this study, we introduce a novel system for joint learning between human operators and robots, that enables human operators to share control of a robot end-effector with a learned assistive agent, simplifies the data collection process, and facilitates simultaneous human demonstration collection and robot manipulation training. As data accumulates, the assistive agent gradually learns. Consequently, less human effort and attention are required, enhancing the efficiency of the data collection process. It also allows the human operator to adjust the control ratio to achieve a trade-off between manual and automated control. We conducted experiments in both simulated environments and physical real-world settings. Through user studies and quantitative evaluations, it is evident that the proposed system could enhance data collection efficiency and reduce the need for human adaptation while ensuring the collected data is of sufficient quality for downstream tasks. \textit{For more details, please refer to our webpage https://norweig1an.github.io/HAJL.github.io/.

comment: 8 pages, 6 figures

♻ ☆ Collaborative Goal Tracking of Multiple Mobile Robots Based on Geometric Graph Neural Network

Weining Lu, Qingquan Lin, Litong Meng, Chenxi Li, Bin Liang

Multiple mobile robots play a significant role in various spatially distributed tasks, highlighting the importance of collaborative path planning to enhance operational efficiency. In unfamiliar and non-repetitive scenarios, reconstructing the global map can be time-inefficient and sometimes unrealistic. Therefore, research has focused on achieving real-time collaborative planning by utilizing sensor data from multiple robots located at different positions, without relying on a global map. This paper introduces a Multi-Robot Collaborative Path Planning method based on a Geometric Graph Neural Network (MRPP-GeoGNN). First, the features of each neighboring robot's sensory data are extracted, and the relative positions of neighboring robots are integrated into each interaction layer to incorporate obstacle information along with location details. Subsequently, GeoGNN maps the amalgamated local environment features to multiple forward directions for the robot's actual movement. An expert data generation method is devised for the robot to advance step by step in the physical environment, generating different expert data in ROS to train the network. We conducted both simulations and physical experiments to validate the effectiveness of the proposed method. Simulation results demonstrate approximately a 5% improvement in accuracy compared to the model based solely on CNN using expert datasets. In the ROS simulation test, the success rate is enhanced by about 4% compared to CNN, and the flow time increase is reduced by approximately 8%, surpassing other GNN models. The physical experimental results indicate that the proposed method enables the robot to navigate successfully in the actual environment and achieve the shortest average path length compared to the benchmark method.

♻ ☆ Optimizing BioTac Simulation for Realistic Tactile Perception IJCNN

Wadhah Zai El Amri, Nicolás Navarro-Guerrero

Tactile sensing presents a promising opportunity for enhancing the interaction capabilities of today's robots. BioTac is a commonly used tactile sensor that enables robots to perceive and respond to physical tactile stimuli. However, the sensor's non-linearity poses challenges in simulating its behavior. In this paper, we first investigate a BioTac simulation that uses temperature, force, and contact point positions to predict the sensor outputs. We show that training with BioTac temperature readings does not yield accurate sensor output predictions during deployment. Consequently, we tested three alternative models, i.e., an XGBoost regressor, a neural network, and a transformer encoder. We train these models without temperature readings and provide a detailed investigation of the window size of the input vectors. We demonstrate that we achieve statistically significant improvements over the baseline network. Furthermore, our results reveal that the XGBoost regressor and transformer outperform traditional feed-forward neural networks in this task. We make all our code and results available online on https://github.com/wzaielamri/Optimizing_BioTac_Simulation.

comment: 12 pages (including appendix), Accepted at the International Joint Conference on Neural Network (IJCNN) 2024, Yokohama, Japan. \c{opyright} 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media... (We refer to IEEE Copyrights)

♻ ☆ OmniRace: 6D Hand Pose Estimation for Intuitive Guidance of Racing Drone

Valerii Serpiva, Aleksey Fedoseev, Sausar Karaf, Ali Alridha Abdulkarim, Dzmitry Tsetserukou

This paper presents the OmniRace approach to controlling a racing drone with 6-degree of freedom (DoF) hand pose estimation and gesture recognition. To our knowledge, it is the first-ever technology that allows for low-level control of high-speed drones using gestures. OmniRace employs a gesture interface based on computer vision and a deep neural network to estimate a 6-DoF hand pose. The advanced machine learning algorithm robustly interprets human gestures, allowing users to control drone motion intuitively. Real-time control of a racing drone demonstrates the effectiveness of the system, validating its potential to revolutionize drone racing and other applications. Experimental results conducted in the Gazebo simulation environment revealed that OmniRace allows the users to complite the UAV race track significantly (by 25.1%) faster and to decrease the length of the test drone path (from 102.9 to 83.7 m). Users preferred the gesture interface for attractiveness (1.57 UEQ score), hedonic quality (1.56 UEQ score), and lower perceived temporal demand (32.0 score in NASA-TLX), while noting the high efficiency (0.75 UEQ score) and low physical demand (19.0 score in NASA-TLX) of the baseline remote controller. The deep neural network attains an average accuracy of 99.75% when applied to both normalized datasets and raw datasets. OmniRace can potentially change the way humans interact with and navigate racing drones in dynamic and complex environments. The source code is available at https://github.com/SerValera/OmniRace.git.

♻ ☆ SLR: Learning Quadruped Locomotion without Privileged Information

Shiyi Chen, Zeyu Wan, Shiyang Yan, Chun Zhang, Weiyi Zhang, Qiang Li, Debing Zhang, Fasih Ud Din Farrukh

The recent mainstream reinforcement learning control for quadruped robots often relies on privileged information, demanding meticulous selection and precise estimation, thereby imposing constraints on the development process. This work proposes a Self-learning Latent Representation (SLR) method, which achieves high-performance control policy learning without the need for privileged information. To enhance the credibility of the proposed method's evaluation, SLR was directly compared with state-of-the-art algorithms using their open-source code repositories and original configuration parameters. Remarkably, SLR surpasses the performance of previous methods using only limited proprioceptive data, demonstrating significant potential for future applications. Ultimately, the trained policy and encoder empower the quadruped robot to traverse various challenging terrains. Videos of our results can be found on our website: https://11chens.github.io/SLR/

♻ ☆ Bimanual Deformable Bag Manipulation Using a Structure-of-Interest Based Neural Dynamics Model

Peng Zhou, Pai Zheng, Jiaming Qi, Chenxi Li, Samantha Lee, Chenguang Yang, David Navarro-Alarcon, Jia Pan

The manipulation of deformable objects by robotic systems presents a significant challenge due to their complex and infinite-dimensional configuration spaces. This paper introduces a novel approach to Deformable Object Manipulation (DOM) by emphasizing the identification and manipulation of Structures of Interest (SOIs) in deformable fabric bags. We propose a bimanual manipulation framework that leverages a Graph Neural Network (GNN)-based latent dynamics model to succinctly represent and predict the behavior of these SOIs. Our approach involves constructing a graph representation from partial point cloud data of the object and learning the latent dynamics model that effectively captures the essential deformations of the fabric bag within a reduced computational space. By integrating this latent dynamics model with Model Predictive Control (MPC), we empower robotic manipulators to perform precise and stable manipulation tasks focused on the SOIs. We have validated our framework through various empirical experiments demonstrating its efficacy in bimanual manipulation of fabric bags. Our contributions not only address the complexities inherent in DOM but also provide new perspectives and methodologies for enhancing robotic interactions with deformable objects by concentrating on their critical structural elements. Experimental videos can be obtained from https://sites.google.com/view/bagbot.

♻ ☆ UADA3D: Unsupervised Adversarial Domain Adaptation for 3D Object Detection with Sparse LiDAR and Large Domain Gaps

Maciej K Wozniak, Mattias Hansson, Marko Thiel, Patric Jensfelt

In this study, we address a gap in existing unsupervised domain adaptation approaches on LiDAR-based 3D object detection, which have predominantly concentrated on adapting between established, high-density autonomous driving datasets. We focus on sparser point clouds, capturing scenarios from different perspectives: not just from vehicles on the road but also from mobile robots on sidewalks, which encounter significantly different environmental conditions and sensor configurations. We introduce Unsupervised Adversarial Domain Adaptation for 3D Object Detection (UADA3D). UADA3D does not depend on pre-trained source models or teacher-student architectures. Instead, it uses an adversarial approach to directly learn domain-invariant features. We demonstrate its efficacy in various adaptation scenarios, showing significant improvements in both self-driving car and mobile robot domains. Our code is open-source and will be available soon.

comment: Accepted for IEEE RA-L 2024

♻ ☆ The Art of Imitation: Learning Long-Horizon Manipulation Tasks from Few Demonstrations

Jan Ole von Hartz, Tim Welschehold, Abhinav Valada, Joschka Boedecker

Task Parametrized Gaussian Mixture Models (TP-GMM) are a sample-efficient method for learning object-centric robot manipulation tasks. However, there are several open challenges to applying TP-GMMs in the wild. In this work, we tackle three crucial challenges synergistically. First, end-effector velocities are non-Euclidean and thus hard to model using standard GMMs. We thus propose to factorize the robot's end-effector velocity into its direction and magnitude, and model them using Riemannian GMMs. Second, we leverage the factorized velocities to segment and sequence skills from complex demonstration trajectories. Through the segmentation, we further align skill trajectories and hence leverage time as a powerful inductive bias. Third, we present a method to automatically detect relevant task parameters per skill from visual observations. Our approach enables learning complex manipulation tasks from just five demonstrations while using only RGB-D observations. Extensive experimental evaluations on RLBench demonstrate that our approach achieves state-of-the-art performance with 20-fold improved sample efficiency. Our policies generalize across different environments, object instances, and object positions, while the learned skills are reusable.

♻ ☆ Flow Matching Imitation Learning for Multi-Support Manipulation RAS 23

Quentin Rouxel, Andrea Ferrari, Serena Ivaldi, Jean-Baptiste Mouret

Humanoid robots could benefit from using their upper bodies for support contacts, enhancing their workspace, stability, and ability to perform contact-rich and pushing tasks. In this paper, we propose a unified approach that combines an optimization-based multi-contact whole-body controller with Flow Matching, a recently introduced method capable of generating multi-modal trajectory distributions for imitation learning. In simulation, we show that Flow Matching is more appropriate for robotics than Diffusion and traditional behavior cloning. On a real full-size humanoid robot (Talos), we demonstrate that our approach can learn a whole-body non-prehensile box-pushing task and that the robot can close dishwasher drawers by adding contacts with its free hand when needed for balance. We also introduce a shared autonomy mode for assisted teleoperation, providing automatic contact placement for tasks not covered in the demonstrations. Full experimental videos are available at: https://hucebot.github.io/flow_multisupport_website/

comment: 2024 IEEE-RAS 23rd International Conference on Humanoid Robots (Humanoids), Nov 2024, Nancy, France

♻ ☆ ATI-CTLO:Adaptive Temporal Interval-based Continuous-Time LiDAR-Only Odometry

Bo Zhou, Jiajie Wu, Yan Pan, Chuanzhao Lu

The motion distortion in LiDAR scans caused by aggressive robot motion and varying terrain features significantly impacts the positioning and mapping performance of 3D LiDAR odometry. Existing distortion correction solutions often struggle to balance computational complexity and accuracy. In this work, we propose an Adaptive Temporal Interval-based Continuous-Time LiDAR-only Odometry, utilizing straightforward and efficient linear interpolation. Our method flexibly adjusts the temporal intervals between control nodes according to the dynamics of motion and environmental characteristics. This adaptability enhances performance across various motion states and improves robustness in challenging, feature-sparse environments. We validate the effectiveness of our method on multiple datasets across different platforms, achieving accuracy comparable to state-of-the-art LiDAR-only odometry methods. Notably, in scenarios involving aggressive motion and sparse features, our method outperforms existing solutions.

♻ ☆ HiRT: Enhancing Robotic Control with Hierarchical Robot Transformers

Jianke Zhang, Yanjiang Guo, Xiaoyu Chen, Yen-Jen Wang, Yucheng Hu, Chengming Shi, Jianyu Chen

Large Vision-Language-Action (VLA) models, leveraging powerful pre trained Vision-Language Models (VLMs) backends, have shown promise in robotic control due to their impressive generalization ability. However, the success comes at a cost. Their reliance on VLM backends with billions of parameters leads to high computational costs and inference latency, limiting the testing scenarios to mainly quasi-static tasks and hindering performance in dynamic tasks requiring rapid interactions. To address these limitations, this paper proposes HiRT, a Hierarchical Robot Transformer framework that enables flexible frequency and performance trade-off. HiRT keeps VLMs running at low frequencies to capture temporarily invariant features while enabling real-time interaction through a high-frequency vision-based policy guided by the slowly updated features. Experiment results in both simulation and real-world settings demonstrate significant improvements over baseline methods. Empirically, in static tasks, we double the control frequency and achieve comparable success rates. Additionally, on novel real-world dynamic ma nipulation tasks which are challenging for previous VLA models, HiRT improves the success rate from 48% to 75%.

♻ ☆ A New Framework for Nonlinear Kalman Filters

Shida Jiang, Junzhe Shi, Scott Moura

The Kalman filter (KF) is a state estimation algorithm that optimally combines system knowledge and measurements to minimize the mean squared error of the estimated states. While KF was initially designed for linear systems, numerous extensions of it, such as extended Kalman filter (EKF), unscented Kalman filter (UKF), cubature Kalman filter (CKF), etc., have been proposed for nonlinear systems. Although different types of nonlinear KFs have different pros and cons, they all use the same framework of linear KF, which, according to what we found in this paper, tends to give overconfident and less accurate state estimations when the measurement functions are nonlinear. Therefore, in this study, we designed a new framework for nonlinear KFs and showed theoretically and empirically that the new framework estimates the states and covariance matrix more accurately than the old one. The new framework was tested on four different nonlinear KFs and five different tasks, showcasing its ability to reduce the estimation errors by several orders of magnitude in low-measurement-noise conditions, with only about a 10 to 90% increase in computational time. All types of nonlinear KFs can benefit from the new framework, and the benefit will increase as the sensors become more and more accurate in the future. As an example, EKF, the simplest nonlinear KF that was previously believed to work poorly for strongly nonlinear systems, can now provide fast and fairly accurate state estimations with the help of the new framework. The codes are available at https://github.com/Shida-Jiang/A-new-framework-for-nonlinear-Kalman-filters.

comment: Some typo fixed

♻ ☆ LiteVLoc: Map-Lite Visual Localization for Image Goal Navigation

Jianhao Jiao, Jinhao He, Changkun Liu, Sebastian Aegidius, Xiangcheng Hu, Tristan Braud, Dimitrios Kanoulas

This paper presents LiteVLoc, a hierarchical visual localization framework that uses a lightweight topo-metric map to represent the environment. The method consists of three sequential modules that estimate camera poses in a coarse-to-fine manner. Unlike mainstream approaches relying on detailed 3D representations, LiteVLoc reduces storage overhead by leveraging learning-based feature matching and geometric solvers for metric pose estimation. A novel dataset for the map-free relocalization task is also introduced. Extensive experiments including localization and navigation in both simulated and real-world scenarios have validate the system's performance and demonstrated its precision and efficiency for large-scale deployment. Code and data will be made publicly available.

comment: 9 pages, 4 figures

♻ ☆ Cardiac Copilot: Automatic Probe Guidance for Echocardiography with World Model MICCAI2024

Haojun Jiang, Zhenguo Sun, Ning Jia, Meng Li, Yu Sun, Shaqi Luo, Shiji Song, Gao Huang

Echocardiography is the only technique capable of real-time imaging of the heart and is vital for diagnosing the majority of cardiac diseases. However, there is a severe shortage of experienced cardiac sonographers, due to the heart's complex structure and significant operational challenges. To mitigate this situation, we present a Cardiac Copilot system capable of providing real-time probe movement guidance to assist less experienced sonographers in conducting freehand echocardiography. This system can enable non-experts, especially in primary departments and medically underserved areas, to perform cardiac ultrasound examinations, potentially improving global healthcare delivery. The core innovation lies in proposing a data-driven world model, named Cardiac Dreamer, for representing cardiac spatial structures. This world model can provide structure features of any cardiac planes around the current probe position in the latent space, serving as an precise navigation map for autonomous plane localization. We train our model with real-world ultrasound data and corresponding probe motion from 110 routine clinical scans with 151K sample pairs by three certified sonographers. Evaluations on three standard planes with 37K sample pairs demonstrate that the world model can reduce navigation errors by up to 33\% and exhibit more stable performance.

comment: Accepted by MICCAI2024

♻ ☆ Data-Driven Dynamics Modeling of Miniature Robotic Blimps Using Neural ODEs With Parameter Auto-Tuning

Yongjian Zhu, Hao Cheng, Feitian Zhang

Miniature robotic blimps, as one type of lighter-than-air aerial vehicles, have attracted increasing attention in the science and engineering community for their enhanced safety, extended endurance, and quieter operation compared to quadrotors. Accurately modeling the dynamics of these robotic blimps poses a significant challenge due to the complex aerodynamics stemming from their large lifting bodies. Traditional first-principle models have difficulty obtaining accurate aerodynamic parameters and often overlook high-order nonlinearities, thus coming to its limit in modeling the motion dynamics of miniature robotic blimps. To tackle this challenge, this letter proposes the Auto-tuning Blimp-oriented Neural Ordinary Differential Equation method (ABNODE), a data-driven approach that integrates first-principle and neural network modeling. Spiraling motion experiments of robotic blimps are conducted, comparing the ABNODE with first-principle and other data-driven benchmark models, the results of which demonstrate the effectiveness of the proposed method.

comment: 8 pages, 8 figures

♻ ☆ Trust or Bust: Ensuring Trustworthiness in Autonomous Weapon Systems

Kasper Cools, Clara Maathuis

The integration of Autonomous Weapon Systems (AWS) into military operations presents both significant opportunities and challenges. This paper explores the multifaceted nature of trust in AWS, emphasising the necessity of establishing reliable and transparent systems to mitigate risks associated with bias, operational failures, and accountability. Despite advancements in Artificial Intelligence (AI), the trustworthiness of these systems, especially in high-stakes military applications, remains a critical issue. Through a systematic review of existing literature, this research identifies gaps in the understanding of trust dynamics during the development and deployment phases of AWS. It advocates for a collaborative approach that includes technologists, ethicists, and military strategists to address these ongoing challenges. The findings underscore the importance of Human-Machine teaming and enhancing system intelligibility to ensure accountability and adherence to International Humanitarian Law. Ultimately, this paper aims to contribute to the ongoing discourse on the ethical implications of AWS and the imperative for trustworthy AI in defense contexts.

comment: Accepted as a workshop paper at MILCOM 2024, 8 pages

♻ ☆ Uncovering the Secrets of Human-Like Movement: A Fresh Perspective on Motion Planning

Lei Shi, Qichao Liu, Cheng Zhou, Wentao Gao, Haotian Wu, Yu Zheng, Xiong Li

This article explores human-like movement from a fresh perspective on motion planning. We analyze the coordinated and compliant movement mechanisms of the human body from the perspective of biomechanics. Based on these mechanisms, we propose an optimal control framework that integrates compliant control dynamics, optimizing robotic arm motion through a response time matrix. This matrix sets the timing parameters for joint movements, turning the system into a time-parameterized optimal control problem. The model focuses on the interaction between active and passive joints under external disturbances, improving adaptability and compliance. This method achieves optimal trajectory generation and balances precision and compliance. Experimental results on both a manipulator and a humanoid robot validate the approach.

comment: 7 pages

♻ ☆ MAL: Motion-Aware Loss with Temporal and Distillation Hints for Self-Supervised Depth Estimation ICRA 2024

Yue-Jiang Dong, Fang-Lue Zhang, Song-Hai Zhang

Depth perception is crucial for a wide range of robotic applications. Multi-frame self-supervised depth estimation methods have gained research interest due to their ability to leverage large-scale, unlabeled real-world data. However, the self-supervised methods often rely on the assumption of a static scene and their performance tends to degrade in dynamic environments. To address this issue, we present Motion-Aware Loss, which leverages the temporal relation among consecutive input frames and a novel distillation scheme between the teacher and student networks in the multi-frame self-supervised depth estimation methods. Specifically, we associate the spatial locations of moving objects with the temporal order of input frames to eliminate errors induced by object motion. Meanwhile, we enhance the original distillation scheme in multi-frame methods to better exploit the knowledge from a teacher network. MAL is a novel, plug-and-play module designed for seamless integration into multi-frame self-supervised monocular depth estimation methods. Adding MAL into previous state-of-the-art methods leads to a reduction in depth estimation errors by up to 4.2% and 10.8% on KITTI and CityScapes benchmarks, respectively.

comment: Accepted by ICRA 2024; Project homepage: https://yuejiangdong.github.io/MotionAwareLoss/

♻ ☆ Behavior-Inspired Neural Networks for Relational Inference

Yulong Yang, Bowen Feng, Keqin Wang, Naomi Leonard, Adji Bousso Dieng, Christine Allen-Blanchette

From pedestrians to Kuramoto oscillators, interactions between agents govern how a multitude of dynamical systems evolve in space and time. Discovering how these agents relate to each other can improve our understanding of the often complex dynamics that underlie these systems. Recent works learn to categorize relationships between agents based on observations of their physical behavior. These approaches are limited in that the relationship categories are modelled as outcomes of categorical distribution, when in real world systems categories often intermingle and interact. In this work, we introduce a level of abstraction between the observable behavior of agents and the latent categories that determine their behavior. To do this, we learn a mapping from agent behavior to agent preferences for each latent category in a graph neural network. We integrate the physical proximity of agents and their preferences in a nonlinear opinion dynamics model which provides a mechanism to identify mutually exclusive latent categories, predict an agent's evolution in time, and control an agent's physical behavior. We demonstrate the utility of our model for learning interpretable categories, and its efficacy on long-horizon prediction across several benchmarks where we outperform existing methods.

☆ Agent-to-Sim: Learning Interactive Behavior Models from Casual Longitudinal Videos

Gengshan Yang, Andrea Bajcsy, Shunsuke Saito, Angjoo Kanazawa

We present Agent-to-Sim (ATS), a framework for learning interactive behavior models of 3D agents from casual longitudinal video collections. Different from prior works that rely on marker-based tracking and multiview cameras, ATS learns natural behaviors of animal and human agents non-invasively through video observations recorded over a long time-span (e.g., a month) in a single environment. Modeling 3D behavior of an agent requires persistent 3D tracking (e.g., knowing which point corresponds to which) over a long time period. To obtain such data, we develop a coarse-to-fine registration method that tracks the agent and the camera over time through a canonical 3D space, resulting in a complete and persistent spacetime 4D representation. We then train a generative model of agent behaviors using paired data of perception and motion of an agent queried from the 4D reconstruction. ATS enables real-to-sim transfer from video recordings of an agent to an interactive behavior simulator. We demonstrate results on pets (e.g., cat, dog, bunny) and human given monocular RGBD videos captured by a smartphone.

comment: Project page: https://gengshan-y.github.io/agent2sim-www/

☆ TexPro: Text-guided PBR Texturing with Procedural Material Modeling

Ziqiang Dang, Wenqi Dong, Zesong Yang, Bangbang Yang, Liang Li, Yuewen Ma, Zhaopeng Cui

In this paper, we present TexPro, a novel method for high-fidelity material generation for input 3D meshes given text prompts. Unlike existing text-conditioned texture generation methods that typically generate RGB textures with baked lighting, TexPro is able to produce diverse texture maps via procedural material modeling, which enables physical-based rendering, relighting, and additional benefits inherent to procedural materials. Specifically, we first generate multi-view reference images given the input textual prompt by employing the latest text-to-image model. We then derive texture maps through a rendering-based optimization with recent differentiable procedural materials. To this end, we design several techniques to handle the misalignment between the generated multi-view images and 3D meshes, and introduce a novel material agent that enhances material classification and matching by exploring both part-level understanding and object-aware material reasoning. Experiments demonstrate the superiority of the proposed method over existing SOTAs and its capability of relighting.

comment: In submission. Supplementary material is included at the end of the main paper (5 pages, 2 figures)

☆ Learning to Synthesize Graphics Programs for Geometric Artworks ICPR 2024

Qi Bing, Chaoyi Zhang, Weidong Cai

Creating and understanding art has long been a hallmark of human ability. When presented with finished digital artwork, professional graphic artists can intuitively deconstruct and replicate it using various drawing tools, such as the line tool, paint bucket, and layer features, including opacity and blending modes. While most recent research in this field has focused on art generation, proposing a range of methods, these often rely on the concept of artwork being represented as a final image. To bridge the gap between pixel-level results and the actual drawing process, we present an approach that treats a set of drawing tools as executable programs. This method predicts a sequence of steps to achieve the final image, allowing for understandable and resolution-independent reproductions under the usage of a set of drawing commands. Our experiments demonstrate that our program synthesizer, Art2Prog, can comprehensively understand complex input images and reproduce them using high-quality executable programs. The experimental results evidence the potential of machines to grasp higher-level information from images and generate compact program-level descriptions.

comment: ICPR 2024

☆ Fully Explicit Dynamic Gaussian Splatting NeurIPS 2024

Junoh Lee, Chang-Yeon Won, Hyunjun Jung, Inhwan Bae, Hae-Gon Jeon

3D Gaussian Splatting has shown fast and high-quality rendering results in static scenes by leveraging dense 3D prior and explicit representations. Unfortunately, the benefits of the prior and representation do not involve novel view synthesis for dynamic motions. Ironically, this is because the main barrier is the reliance on them, which requires increasing training and rendering times to account for dynamic motions. In this paper, we design a Explicit 4D Gaussian Splatting(Ex4DGS). Our key idea is to firstly separate static and dynamic Gaussians during training, and to explicitly sample positions and rotations of the dynamic Gaussians at sparse timestamps. The sampled positions and rotations are then interpolated to represent both spatially and temporally continuous motions of objects in dynamic scenes as well as reducing computational cost. Additionally, we introduce a progressive training scheme and a point-backtracking technique that improves Ex4DGS's convergence. We initially train Ex4DGS using short timestamps and progressively extend timestamps, which makes it work well with a few point clouds. The point-backtracking is used to quantify the cumulative error of each Gaussian over time, enabling the detection and removal of erroneous Gaussians in dynamic scenes. Comprehensive experiments on various scenes demonstrate the state-of-the-art rendering quality from our method, achieving fast rendering of 62 fps on a single 2080Ti GPU.

comment: Accepted at NeurIPS 2024

★ Deep Learning and Machine Learning -- Object Detection and Semantic Segmentation: From Theory to Applications

Jintao Ren, Ziqian Bi, Qian Niu, Junyu Liu, Benji Peng, Sen Zhang, Xuanhe Pan, Jinlang Wang, Keyu Chen, Caitlyn Heqi Yin, Pohsun Feng, Yizhu Wen, Tianyang Wang, Silin Chen, Ming Li, Jiawei Xu, Ming Liu

This book offers an in-depth exploration of object detection and semantic segmentation, combining theoretical foundations with practical applications. It covers state-of-the-art advancements in machine learning and deep learning, with a focus on convolutional neural networks (CNNs), YOLO architectures, and transformer-based approaches like DETR. The book also delves into the integration of artificial intelligence (AI) techniques and large language models for enhanced object detection in complex environments. A thorough discussion of big data analysis is presented, highlighting the importance of data processing, model optimization, and performance evaluation metrics. By bridging the gap between traditional methods and modern deep learning frameworks, this book serves as a comprehensive guide for researchers, data scientists, and engineers aiming to leverage AI-driven methodologies in large-scale object detection tasks.

comment: 167 pages

♻ ☆ Efficient Nearest Neighbor Search Using Dynamic Programming

Pengfei Wang, Jiantao Song, Shiqing Xin, Shuangmin Chen, Changhe Tu, Wenping Wang, Jiaye Wang

Given a collection of points in R^3, KD-Tree and R-Tree are well-known nearest neighbor search (NNS) algorithms that rely on space partitioning and spatial indexing techniques. However, when the query point is far from the data points or the data points inherently represent a 2-manifold surface, their query performance may degrade. To address this, we propose a novel dynamic programming technique that precomputes a Directed Acyclic Graph (DAG) to encode the proximity structure between data points. More specifically, the DAG captures how the proximity structure evolves during the incremental construction of the Voronoi diagram of the data points. Experimental results demonstrate that our method achieves a 1x-10x speedup. Additionally, our algorithm offers several valuable features. For instance, it naturally supports an O(k \log n) algorithm for farthest point sampling, where k is the desired number of sample points. Moreover, density peak clustering, which involves finding the nearest point among the top K points, is typically considered to have a time complexity of O(n^2). With our algorithm, this can be reduced to O(n \log n). We believe this work will inspire further research on the NNS problem.

♻ ☆ Inverse Garment and Pattern Modeling with a Differentiable Simulator

Boyang Yu, Frederic Cordier, Hyewon Seo

The capability to generate simulation-ready garment models from 3D shapes of clothed humans will significantly enhance the interpretability of captured geometry of real garments, as well as their faithful reproduction in the virtual world. This will have notable impact on fields like shape capture in social VR, and virtual try-on in the fashion industry. To align with the garment modeling process standardized by the fashion industry as well as cloth simulation softwares, it is required to recover 2D patterns. This involves an inverse garment design problem, which is the focus of our work here: Starting with an arbitrary target garment geometry, our system estimates an animatable garment model by automatically adjusting its corresponding 2D template pattern, along with the material parameters of the physics-based simulation (PBS). Built upon a differentiable cloth simulator, the optimization process is directed towards minimizing the deviation of the simulated garment shape from the target geometry. Moreover, our produced patterns meet manufacturing requirements such as left-to-right-symmetry, making them suited for reverse garment fabrication. We validate our approach on examples of different garment types, and show that our method faithfully reproduces both the draped garment shape and the sewing pattern.

♻ ☆ End-to-End Rate-Distortion Optimized 3D Gaussian Representation ECCV 2024

Henan Wang, Hanxin Zhu, Tianyu He, Runsen Feng, Jiajun Deng, Jiang Bian, Zhibo Chen

3D Gaussian Splatting (3DGS) has become an emerging technique with remarkable potential in 3D representation and image rendering. However, the substantial storage overhead of 3DGS significantly impedes its practical applications. In this work, we formulate the compact 3D Gaussian learning as an end-to-end Rate-Distortion Optimization (RDO) problem and propose RDO-Gaussian that can achieve flexible and continuous rate control. RDO-Gaussian addresses two main issues that exist in current schemes: 1) Different from prior endeavors that minimize the rate under the fixed distortion, we introduce dynamic pruning and entropy-constrained vector quantization (ECVQ) that optimize the rate and distortion at the same time. 2) Previous works treat the colors of each Gaussian equally, while we model the colors of different regions and materials with learnable numbers of parameters. We verify our method on both real and synthetic scenes, showcasing that RDO-Gaussian greatly reduces the size of 3D Gaussian over 40x, and surpasses existing methods in rate-distortion performance.

comment: ECCV 2024

☆ GRS: Generating Robotic Simulation Tasks from Real-World Images

Alex Zook, Fan-Yun Sun, Josef Spjut, Valts Blukis, Stan Birchfield, Jonathan Tremblay

We introduce GRS (Generating Robotic Simulation tasks), a novel system to address the challenge of real-to-sim in robotics, computer vision, and AR/VR. GRS enables the creation of digital twin simulations from single real-world RGB-D observations, complete with diverse, solvable tasks for virtual agent training. We use state-of-the-art vision-language models (VLMs) to achieve a comprehensive real-to-sim pipeline. GRS operates in three stages: 1) scene comprehension using SAM2 for object segmentation and VLMs for object description, 2) matching identified objects with simulation-ready assets, and 3) generating contextually appropriate robotic tasks. Our approach ensures simulations align with task specifications by generating test suites designed to verify adherence to the task specification. We introduce a router that iteratively refines the simulation and test code to ensure the simulation is solvable by a robot policy while remaining aligned to the task specification. Our experiments demonstrate the system's efficacy in accurately identifying object correspondence, which allows us to generate task environments that closely match input environments, and enhance automated simulation task generation through our novel router mechanism.

☆ Quasi-Static Continuum Model of Octopus-Like Soft Robot Arm Under Water Actuated by Twisted and Coiled Artificial Muscles (TCAMs)

Amirreza Fahim Golestaneh, Venanzio Cichella, Caterina Lamuta

The current work is a qualitative study that aims to explore the implementation of Twisted and Coiled Artificial Muscles (TCAMs) for actuating and replicating the bending motion of an octopus-like soft robot arm underwater. Additionally, it investigates the impact of hydrostatic and dynamic forces from steady-state fluid flow on the arm's motion. The artificial muscles are lightweight and low-cost actuators that generate a high power-to-weight ratio, producing tensile force up to 12,600 times their own weight, which is close to the functionality of biological muscles. The "extended" Cosserat theory of rods is employed to formulate a quasi-static continuum model of arm motion, where the arm's cross-section is not only capable of rigid rotation but also deforms within its plane. This planar deformation of the arm cross-section aligns with the biological behavior of the octopus arm, where the stiffness of the hydrostat is directly induced by the incompressibility of the tissues. In line with the main goal, a constitutive model is derived for the material of the octopus arm to capture its characteristic behavior.

comment: 12 pages, Under review at the journal "Robotics Reports"

☆ Generative AI Agents in Autonomous Machines: A Safety Perspective

Jason Jabbour, Vijay Janapa Reddi

The integration of Generative Artificial Intelligence (AI) into autonomous machines represents a major paradigm shift in how these systems operate and unlocks new solutions to problems once deemed intractable. Although generative AI agents provide unparalleled capabilities, they also have unique safety concerns. These challenges require robust safeguards, especially for autonomous machines that operate in high-stakes environments. This work investigates the evolving safety requirements when generative models are integrated as agents into physical autonomous machines, comparing these to safety considerations in less critical AI applications. We explore the challenges and opportunities to ensure the safe deployment of generative AI-driven autonomous machines. Furthermore, we provide a forward-looking perspective on the future of AI-driven autonomous systems and emphasize the importance of evaluating and communicating safety risks. As an important step towards addressing these concerns, we recommend the development and implementation of comprehensive safety scorecards for the use of generative AI technologies in autonomous machines.

☆ Evaluating Transferable Emotion Expressions for Zoomorphic Social Robots using VR Prototyping

Shaun Macdonald, Robin Bretin, Salma ElSayed

Zoomorphic robots have the potential to offer companionship and well-being as accessible, low-maintenance alternatives to pet ownership. Many such robots, however, feature limited emotional expression, restricting their potential for rich affective relationships with everyday domestic users. Additionally, exploring this design space using hardware prototyping is obstructed by physical and logistical constraints. We leveraged virtual reality rapid prototyping with passive haptic interaction to conduct a broad mixed-methods evaluation of emotion expression modalities and participatory prototyping of multimodal expressions. We found differences in recognisability, effectiveness and user empathy between modalities while highlighting the importance of facial expressions and the benefits of combining animal-like and unambiguous modalities. We use our findings to inform promising directions for the affective zoomorphic robot design and potential implementations via hardware modification or augmented reality, then discuss how VR prototyping makes this field more accessible to designers and researchers.

comment: 10 pages, 9 figures, accepted to 23rd IEEE International Symposium on Mixed and Augmented Reality (ISMAR 2024)

☆ AssemblyComplete: 3D Combinatorial Construction with Deep Reinforcement Learning

Alan Chen, Changliu Liu

A critical goal in robotics and autonomy is to teach robots to adapt to real-world collaborative tasks, particularly in automatic assembly. The ability of a robot to understand the original intent of an incomplete assembly and complete missing features without human instruction is valuable but challenging. This paper introduces 3D combinatorial assembly completion, which is demonstrated using combinatorial unit primitives (i.e., Lego bricks). Combinatorial assembly is challenging due to the possible assembly combinations and complex physical constraints (e.g., no brick collisions, structure stability, inventory constraints, etc.). To address these challenges, we propose a two-part deep reinforcement learning (DRL) framework that tackles teaching the robot to understand the objective of an incomplete assembly and learning a construction policy to complete the assembly. The robot queries a stable object library to facilitate assembly inference and guide learning. In addition to the robot policy, an action mask is developed to rule out invalid actions that violate physical constraints for object-oriented construction. We demonstrate the proposed framework's feasibility and robustness in a variety of assembly scenarios in which the robot satisfies real-life assembly with respect to both solution and runtime quality. Furthermore, results demonstrate that the proposed framework effectively infers and assembles incomplete structures for unseen and unique object types.

comment: Submitted to 2025 American Control Conference (ACC)

☆ EVA: An Embodied World Model for Future Video Anticipation

Xiaowei Chi, Hengyuan Zhang, Chun-Kai Fan, Xingqun Qi, Rongyu Zhang, Anthony Chen, Chi-min Chan, Wei Xue, Wenhan Luo, Shanghang Zhang, Yike Guo

World models integrate raw data from various modalities, such as images and language to simulate comprehensive interactions in the world, thereby displaying crucial roles in fields like mixed reality and robotics. Yet, applying the world model for accurate video prediction is quite challenging due to the complex and dynamic intentions of the various scenes in practice. In this paper, inspired by the human rethinking process, we decompose the complex video prediction into four meta-tasks that enable the world model to handle this issue in a more fine-grained manner. Alongside these tasks, we introduce a new benchmark named Embodied Video Anticipation Benchmark (EVA-Bench) to provide a well-rounded evaluation. EVA-Bench focused on evaluating the video prediction ability of human and robot actions, presenting significant challenges for both the language model and the generation model. Targeting embodied video prediction, we propose the Embodied Video Anticipator (EVA), a unified framework aiming at video understanding and generation. EVA integrates a video generation model with a visual language model, effectively combining reasoning capabilities with high-quality generation. Moreover, to enhance the generalization of our framework, we tailor-designed a multi-stage pretraining paradigm that adaptatively ensembles LoRA to produce high-fidelity results. Extensive experiments on EVA-Bench highlight the potential of EVA to significantly improve performance in embodied scenes, paving the way for large-scale pre-trained models in real-world prediction tasks.

☆ Lie Theory Based Optimization for Unified State Planning of Mobile Manipulators

William Smith, Siddharth Singh, Julia Rudy, Yuxiang Guan

Mobile manipulators are finding use in numerous practical applications. The current issues with mobile manipulation are the large state space owing to the mobile base and the challenge of modeling high degree of freedom systems. It is critical to devise fast and accurate algorithms that generate smooth motion plans for such mobile manipulators. Existing techniques attempt to solve this problem but focus on separating the motion of the base and manipulator. We propose an approach using Lie theory to find the inverse kinematic constraints by converting the kinematic model, created using screw coordinates, between its Lie group and vector representation. An optimization function is devised to solve for the desired joint states of the entire mobile manipulator. This allows the motion of the mobile base and manipulator to be planned and applied in unison resulting in a smooth and accurate motion plan. The performance of the proposed state planner is validated on simulated mobile manipulators in an analytical experiment. Our solver is available with further derivations and results at https://github.com/peleito/slithers.

comment: 8 pages, 9 figures, conference submission

☆ An Agile Large-Workspace Teleoperation Interface Based on Human Arm Motion and Force Estimation

Jianhang Jia, Hao Zhou, Xin Zhang

Teleoperation can transfer human perception and cognition to a slave robot to cope with some complex tasks, in which the agility and flexibility of the interface play an important role in mapping human intention to the robot. In this paper, we developed an agile large-workspace teleoperation interface by estimating human arm behavior. Using the wearable sensor, namely the inertial measurement unit and surface electromyography armband, we can capture the human arm motion and force information, thereby intuitively controlling the manipulation of the robot. The control principle of our wearable interface includes two parts: (1) the arm incremental kinematics and (2) the grasping recognition. Moreover, we developed a teleoperation framework with a time synchronization mechanism for the real-time application. We conducted experimental comparisons with a versatile haptic device (Omega 7) to verify the effectiveness of our interface and framework. Seven subjects are invited to complete three different tasks: free motion, handover, and pick-and-place action (each task ten times), and the total number of tests is 420. Objectively, we used the task completion time and success rate to compare the performance of the two interfaces quantitatively. In addition, to quantify the operator experience, we used the NASA Task Load Index to assess their subjective feelings. The results showed that the proposed interface achieved a competitive performance with a better operating experience.

comment: 6 pages, 8 figures, accepted by 2024 IEEE International Conference on Robotics and Biomimetics (IEEE ROBIO 2024)

☆ Evaluation of Human-Robot Interfaces based on 2D/3D Visual and Haptic Feedback for Aerial Manipulation

Julien Mellet, Mike Allenspach, Eugenio Cuniato, Claudio Pacchierotti, Roland Siegwart, Marco Tognon

Most telemanipulation systems for aerial robots provide the operator with only 2D screen visual information. The lack of richer information about the robot's status and environment can limit human awareness and, in turn, task performance. While the pilot's experience can often compensate for this reduced flow of information, providing richer feedback is expected to reduce the cognitive workload and offer a more intuitive experience overall. This work aims to understand the significance of providing additional pieces of information during aerial telemanipulation, namely (i) 3D immersive visual feedback about the robot's surroundings through mixed reality (MR) and (ii) 3D haptic feedback about the robot interaction with the environment. To do so, we developed a human-robot interface able to provide this information. First, we demonstrate its potential in a real-world manipulation task requiring sub-centimeter-level accuracy. Then, we evaluate the individual effect of MR vision and haptic feedback on both dexterity and workload through a human subjects study involving a virtual block transportation task. Results show that both 3D MR vision and haptic feedback improve the operator's dexterity in the considered teleoperated aerial interaction tasks. Nevertheless, pilot experience remains the most significant factor.

comment: 12 pages, 11 figures, journal paper

☆ A Semi-decentralized and Variational-Equilibrium-Based Trajectory Planner for Connected and Autonomous Vehicles

Zhengqin Liu, Jinlong Lei, Peng Yi

This paper designs a novel trajectory planning approach to resolve the computational efficiency and safety problems in uncoordinated methods by exploiting vehicle-to-everything (V2X) technology. The trajectory planning for connected and autonomous vehicles (CAVs) is formulated as a game with coupled safety constraints. We then define interaction-fair trajectories and prove that they correspond to the variational equilibrium (VE) of this game. We propose a semi-decentralized planner for the vehicles to seek VE-based fair trajectories, which can significantly improve computational efficiency through parallel computing among CAVs and enhance the safety of planned trajectories by ensuring equilibrium concordance among CAVs. Finally, experimental results show the advantages of the approach, including fast computation speed, high scalability, equilibrium concordance, and safety.

☆ DynaVINS++: Robust Visual-Inertial State Estimator in Dynamic Environments by Adaptive Truncated Least Squares and Stable State Recovery

Seungwon Song, Hyungtae Lim, Alex Junho Lee, Hyun Myung

Despite extensive research in robust visual-inertial navigation systems~(VINS) in dynamic environments, many approaches remain vulnerable to objects that suddenly start moving, which are referred to as \textit{abruptly dynamic objects}. In addition, most approaches have considered the effect of dynamic objects only at the feature association level. In this study, we observed that the state estimation diverges when errors from false correspondences owing to moving objects incorrectly propagate into the IMU bias terms. To overcome these problems, we propose a robust VINS framework called \mbox{\textit{DynaVINS++}}, which employs a) adaptive truncated least square method that adaptively adjusts the truncation range using both feature association and IMU preintegration to effectively minimize the effect of the dynamic objects while reducing the computational cost, and b)~stable state recovery with bias consistency check to correct misestimated IMU bias and to prevent the divergence caused by abruptly dynamic objects. As verified in both public and real-world datasets, our approach shows promising performance in dynamic environments, including scenes with abruptly dynamic objects.

comment: 8 pages, 7 figures. S. Song, H. Lim, A. J. Lee and H. Myung, "DynaVINS++: Robust Visual-Inertial State Estimator in Dynamic Environments by Adaptive Truncated Least Squares and Stable State Recovery," in IEEE Robotics and Automation Letters, vol. 9, no. 10, pp. 9127-9134, Oct. 2024

☆ Integrated Design and Control of a Robotic Arm on a Quadcopter for Enhanced Package Delivery

Animesh Singh, Jason Hillyer, Fariba Ariaei, Hossein Jula

This paper presents a comprehensive design process for the integration of a robotic arm into a quadcopter, emphasizing the physical modeling, system integration, and controller development. Utilizing SolidWorks for mechanical design and MATLAB Simscape for simulation and control, this study addresses the challenges encountered in integrating the robotic arm with the drone, encompassing both mechanical and control aspects. Two types of controllers are developed and analyzed: a Proportional-Integral-Derivative (PID) controller and a Model Reference Adaptive Controller (MRAC). The design and tuning of these controllers are key components of this research, with the focus on their application in package delivery tasks. Extensive simulations demonstrate the performance of each controller, with PID controllers exhibiting superior trajectory tracking and lower Root Mean Square (RMS) errors under various payload conditions. The results underscore the efficacy of PID control for stable flight and precise maneuvering, while highlighting adaptability of MRAC to changing dynamics.

☆ Large Language Models for Autonomous Driving (LLM4AD): Concept, Benchmark, Simulation, and Real-Vehicle Experiment

Can Cui, Yunsheng Ma, Zichong Yang, Yupeng Zhou, Peiran Liu, Juanwu Lu, Lingxi Li, Yaobin Chen, Jitesh H. Panchal, Amr Abdelraouf, Rohit Gupta, Kyungtae Han, Ziran Wang

With the broader usage and highly successful development of Large Language Models (LLMs), there has been a growth of interest and demand for applying LLMs to autonomous driving technology. Driven by their natural language understanding and reasoning ability, LLMs have the potential to enhance various aspects of autonomous driving systems, from perception and scene understanding to language interaction and decision-making. In this paper, we first introduce novel concepts and approaches to designing LLMs for autonomous driving (LLM4AD). Then, we propose a comprehensive benchmark for evaluating the instruction-following abilities of LLMs within the autonomous driving domain. Furthermore, we conduct a series of experiments on both simulation and real-world vehicle platforms, thoroughly evaluating the performance and potential of our LLM4AD systems. Our research highlights the significant potential of LLMs to enhance various aspects of autonomous vehicle technology, from perception and scene understanding to language interaction and decision-making.

☆ An Image-Guided Robotic System for Transcranial Magnetic Stimulation: System Development and Experimental Evaluation

Yihao Liu, Jiaming Zhang, Letian Ai, Jing Tian, Shahriar Sefati, Huan Liu, Alejandro Martin-Gomez, Amir Kheradmand, Mehran Armand

Transcranial magnetic stimulation (TMS) is a noninvasive medical procedure that can modulate brain activity, and it is widely used in neuroscience and neurology research. Compared to manual operators, robots may improve the outcome of TMS due to their superior accuracy and repeatability. However, there has not been a widely accepted standard protocol for performing robotic TMS using fine-segmented brain images, resulting in arbitrary planned angles with respect to the true boundaries of the modulated cortex. Given that the recent study in TMS simulation suggests a noticeable difference in outcomes when using different anatomical details, cortical shape should play a more significant role in deciding the optimal TMS coil pose. In this work, we introduce an image-guided robotic system for TMS that focuses on (1) establishing standardized planning methods and heuristics to define a reference (true zero) for the coil poses and (2) solving the issue that the manual coil placement requires expert hand-eye coordination which often leading to low repeatability of the experiments. To validate the design of our robotic system, a phantom study and a preliminary human subject study were performed. Our results show that the robotic method can half the positional error and improve the rotational accuracy by up to two orders of magnitude. The accuracy is proven to be repeatable because the standard deviation of multiple trials is lowered by an order of magnitude. The improved actuation accuracy successfully translates to the TMS application, with a higher and more stable induced voltage in magnetic field sensors.

comment: This work has been submitted to the IEEE for possible publication

♻ ☆ ORSO: Accelerating Reward Design via Online Reward Selection and Policy Optimization

Chen Bo Calvin Zhang, Zhang-Wei Hong, Aldo Pacchiano, Pulkit Agrawal

Reward shaping is a critical component in reinforcement learning (RL), particularly for complex tasks where sparse rewards can hinder learning. While shaping rewards have been introduced to provide additional guidance, selecting effective shaping functions remains challenging and computationally expensive. This paper introduces Online Reward Selection and Policy Optimization (ORSO), a novel approach that frames shaping reward selection as an online model selection problem. ORSO employs principled exploration strategies to automatically identify promising shaping reward functions without human intervention, balancing exploration and exploitation with provable regret guarantees. We demonstrate ORSO's effectiveness across various continuous control tasks using the Isaac Gym simulator. Compared to traditional methods that fully evaluate each shaping reward function, ORSO significantly improves sample efficiency, reduces computational time, and consistently identifies high-quality reward functions that produce policies comparable to those generated by domain experts through hand-engineered rewards.

comment: preprint, 35 pages, 23 figures

♻ ☆ Robust RL with LLM-Driven Data Synthesis and Policy Adaptation for Autonomous Driving

Sihao Wu, Jiaxu Liu, Xiangyu Yin, Guangliang Cheng, Xingyu Zhao, Meng Fang, Xinping Yi, Xiaowei Huang

The integration of Large Language Models (LLMs) into autonomous driving systems demonstrates strong common sense and reasoning abilities, effectively addressing the pitfalls of purely data-driven methods. Current LLM-based agents require lengthy inference times and face challenges in interacting with real-time autonomous driving environments. A key open question is whether we can effectively leverage the knowledge from LLMs to train an efficient and robust Reinforcement Learning (RL) agent. This paper introduces RAPID, a novel \underline{\textbf{R}}obust \underline{\textbf{A}}daptive \underline{\textbf{P}}olicy \underline{\textbf{I}}nfusion and \underline{\textbf{D}}istillation framework, which trains specialized mix-of-policy RL agents using data synthesized by an LLM-based driving agent and online adaptation. RAPID features three key designs: 1) utilization of offline data collected from an LLM agent to distil expert knowledge into RL policies for faster real-time inference; 2) introduction of robust distillation in RL to inherit both performance and robustness from LLM-based teacher; and 3) employment of a mix-of-policy approach for joint decision decoding with a policy adapter. Through fine-tuning via online environment interaction, RAPID reduces the forgetting of LLM knowledge while maintaining adaptability to different tasks. Extensive experiments demonstrate RAPID's capability to effectively integrate LLM knowledge into scaled-down RL policies in an efficient, adaptable, and robust way. Code and checkpoints will be made publicly available upon acceptance.

♻ ☆ Safety-Critical Formation Control of Non-Holonomic Multi-Robot Systems in Communication-Limited Environments

Vishrut Bohara, Siavash Farzan

This paper presents a novel estimator-based safety-critical controller for formation control of non-holonomic mobile robots in communication-limited environments. The proposed decentralized framework integrates a robust state estimator with a formation tracking control law, addressing the challenges of inter-agent collision avoidance and disturbance attenuation in leader-follower formations using control barrier functions. The estimator's design accounts for both constant and time-varying velocity profiles, enhancing the system's adaptability to dynamic scenarios. A closed-form solution for the tracking controller facilitates efficient implementation while maintaining formation integrity. The incorporation of string stability metrics further reinforces the framework's resilience against propagating disturbances from predecessors. Rigorous stability analysis using Lyapunov functions ensures the stability of estimation errors and the convergence of the formation to desired configurations. The effectiveness and robustness of the proposed approach are validated through numerical simulations of various maneuvers and realistic Gazebo experiments involving formations in a warehouse environment. The results demonstrate the controller's ability to maintain safety, achieve precise formation control, and mitigate disturbances in scenarios without inter-robot communication.

comment: Under review

♻ ★ Octopus: Embodied Vision-Language Programmer from Environmental Feedback

Jingkang Yang, Yuhao Dong, Shuai Liu, Bo Li, Ziyue Wang, Chencheng Jiang, Haoran Tan, Jiamu Kang, Yuanhan Zhang, Kaiyang Zhou, Ziwei Liu

Large vision-language models (VLMs) have achieved substantial progress in multimodal perception and reasoning. When integrated into an embodied agent, existing embodied VLM works either output detailed action sequences at the manipulation level or only provide plans at an abstract level, leaving a gap between high-level planning and real-world manipulation. To bridge this gap, we introduce Octopus, an embodied vision-language programmer that uses executable code generation as a medium to connect planning and manipulation. Octopus is designed to 1) proficiently comprehend an agent's visual and textual task objectives, 2) formulate intricate action sequences, and 3) generate executable code. To facilitate Octopus model development, we introduce OctoVerse: a suite of environments tailored for benchmarking vision-based code generators on a wide spectrum of tasks, ranging from mundane daily chores in simulators to sophisticated interactions in complex video games such as Grand Theft Auto (GTA) and Minecraft. To train Octopus, we leverage GPT-4 to control an explorative agent that generates training data, i.e., action blueprints and corresponding executable code. We also collect feedback that enables an enhanced training scheme called Reinforcement Learning with Environmental Feedback (RLEF). Through a series of experiments, we demonstrate Octopus's functionality and present compelling results, showing that the proposed RLEF refines the agent's decision-making. By open-sourcing our simulation environments, dataset, and model architecture, we aspire to ignite further innovation and foster collaborative applications within the broader embodied AI community.

comment: Project Page: https://choiszt.github.io/Octopus/, Codebase: https://github.com/dongyh20/Octopus

♻ ☆ DTG : Diffusion-based Trajectory Generation for Mapless Global Navigation

Jing Liang, Amirreza Payandeh, Daeun Song, Xuesu Xiao, Dinesh Manocha

We present a novel end-to-end diffusion-based trajectory generation method, DTG, for mapless global navigation in challenging outdoor scenarios with occlusions and unstructured off-road features like grass, buildings, bushes, etc. Given a distant goal, our approach computes a trajectory that satisfies the following goals: (1) minimize the travel distance to the goal; (2) maximize the traversability by choosing paths that do not lie in undesirable areas. Specifically, we present a novel Conditional RNN(CRNN) for diffusion models to efficiently generate trajectories. Furthermore, we propose an adaptive training method that ensures that the diffusion model generates more traversable trajectories. We evaluate our methods in various outdoor scenes and compare the performance with other global navigation algorithms on a Husky robot. In practice, we observe at least a 15% improvement in traveling distance and around a 7% improvement in traversability.

comment: 10 pages

♻ ☆ Rapid and Robust Trajectory Optimization for Humanoids

Bohao Zhang, Ram Vasudevan

Performing trajectory design for humanoid robots with high degrees of freedom is computationally challenging. The trajectory design process also often involves carefully selecting various hyperparameters and requires a good initial guess which can further complicate the development process. This work introduces a generalized gait optimization framework that directly generates smooth and physically feasible trajectories. The proposed method demonstrates faster and more robust convergence than existing techniques and explicitly incorporates closed-loop kinematic constraints that appear in many modern humanoids. The method is implemented as an open-source C++ codebase which can be found at https://roahmlab.github.io/RAPTOR/.

♻ ☆ ViSaRL: Visual Reinforcement Learning Guided by Human Saliency

Anthony Liang, Jesse Thomason, Erdem Bıyık

Training robots to perform complex control tasks from high-dimensional pixel input using reinforcement learning (RL) is sample-inefficient, because image observations are comprised primarily of task-irrelevant information. By contrast, humans are able to visually attend to task-relevant objects and areas. Based on this insight, we introduce Visual Saliency-Guided Reinforcement Learning (ViSaRL). Using ViSaRL to learn visual representations significantly improves the success rate, sample efficiency, and generalization of an RL agent on diverse tasks including DeepMind Control benchmark, robot manipulation in simulation and on a real robot. We present approaches for incorporating saliency into both CNN and Transformer-based encoders. We show that visual representations learned using ViSaRL are robust to various sources of visual perturbations including perceptual noise and scene variations. ViSaRL nearly doubles success rate on the real-robot tasks compared to the baseline which does not use saliency.

♻ ☆ ORLA*: Mobile Manipulator-Based Object Rearrangement with Lazy A Star ICRA 2025

Kai Gao, Zhaxizhuoma, Yan Ding, Shiqi Zhang, Jingjin Yu

Effectively performing object rearrangement is an essential skill for mobile manipulators, e.g., setting up a dinner table or organizing a desk. A key challenge in such problems is deciding an appropriate manipulation order for objects to effectively untangle dependencies between objects while considering the necessary motions for realizing the manipulations (e.g., pick and place). To our knowledge, computing time-optimal multi-object rearrangement solutions for mobile manipulators remains a largely untapped research direction. In this research, we propose ORLA*, which leverages delayed (lazy) evaluation in searching for a high-quality object pick and place sequence that considers both end-effector and mobile robot base travel. ORLA* also supports multi-layered rearrangement tasks considering pile stability using machine learning. Employing an optimal solver for finding temporary locations for displacing objects, ORLA* can achieve global optimality. Through extensive simulation and ablation study, we confirm the effectiveness of ORLA* delivering quality solutions for challenging rearrangement instances. Supplementary materials are available at: https://gaokai15.github.io/ORLA-Star/

comment: Submitted to ICRA 2025

♻ ☆ 3D-BBS: Global Localization for 3D Point Cloud Scan Matching Using Branch-and-Bound Algorithm ICRA2024

Koki Aoki, Kenji Koide, Shuji Oishi, Masashi Yokozuka, Atsuhiko Banno, Junichi Meguro

This paper presents an accurate and fast 3D global localization method, 3D-BBS, that extends the existing branch-and-bound (BnB)-based 2D scan matching (BBS) algorithm. To reduce memory consumption, we utilize a sparse hash table for storing hierarchical 3D voxel maps. To improve the processing cost of BBS in 3D space, we propose an efficient roto-translational space branching. Furthermore, we devise a batched BnB algorithm to fully leverage GPU parallel processing. Through experiments in simulated and real environments, we demonstrated that the 3D-BBS enabled accurate global localization with only a 3D LiDAR scan roughly aligned in the gravity direction and a 3D pre-built map. This method required only 878 msec on average to perform global localization and outperformed state-of-the-art global registration methods in terms of accuracy and processing speed.

comment: IEEE International Conference on Robotics and Automation (ICRA2024)

♻ ☆ Mitigating Side Effects in Multi-Agent Systems Using Blame Assignment

Pulkit Rustagi, Sandhya Saisubramanian

When independently trained or designed robots are deployed in a shared environment, their combined actions can lead to unintended negative side effects (NSEs). To ensure safe and efficient operation, robots must optimize task performance while minimizing the penalties associated with NSEs, balancing individual objectives with collective impact. We model the problem of mitigating NSEs in a cooperative multi-agent system as a bi-objective lexicographic decentralized Markov decision process. We assume independence of transitions and rewards with respect to the robots' tasks, but the joint NSE penalty creates a form of dependence in this setting. To improve scalability, the joint NSE penalty is decomposed into individual penalties for each robot using credit assignment, which facilitates decentralized policy computation. We empirically demonstrate, using mobile robots and in simulation, the effectiveness and scalability of our approach in mitigating NSEs.

comment: 8 pages, 5 figures

♻ ☆ Granger Causal Interaction Skill Chains

Caleb Chuck, Kevin Black, Aditya Arjun, Yuke Zhu, Scott Niekum

Reinforcement Learning (RL) has demonstrated promising results in learning policies for complex tasks, but it often suffers from low sample efficiency and limited transferability. Hierarchical RL (HRL) methods aim to address the difficulty of learning long-horizon tasks by decomposing policies into skills, abstracting states, and reusing skills in new tasks. However, many HRL methods require some initial task success to discover useful skills, which paradoxically may be very unlikely without access to useful skills. On the other hand, reward-free HRL methods often need to learn far too many skills to achieve proper coverage in high-dimensional domains. In contrast, we introduce the Chain of Interaction Skills (COInS) algorithm, which focuses on controllability in factored domains to identify a small number of task-agnostic skills that still permit a high degree of control. COInS uses learned detectors to identify interactions between state factors and then trains a chain of skills to control each of these factors successively. We evaluate COInS on a robotic pushing task with obstacles -- a challenging domain where other RL and HRL methods fall short. We also demonstrate the transferability of skills learned by COInS, using variants of Breakout, a common RL benchmark, and show 2-3x improvement in both sample efficiency and final performance compared to standard RL baselines.

comment: Accepted TMLR 2024